Assembly Standing Committee on Higher Education

Good afternoon, the Assembly - welcome everybody. I'm Assemblymember Mike Fong, Chair of the Assembly Committee on Higher Education. I'd like to welcome all of you to this Joint Hearing on the California State University's AI Empowered Initiative. Whether you're here in person or watching virtually, I would like to thank you for your participation.

Please note that all materials for today's hearing can be accessed online at http://ahed.assembly.ca.gov/hearings. I want to go over some key elements of the structure of today's oversight hearing. We have a few different options for testimony today in the hearing room and the submission of written comments. When we reach the public comment portion of the agenda.

I would ask any Members of the public here in the room to approach the microphone and form a line. Due to time constraints, I'll be limiting public testimony to one minute. You can exit the hearing once you finish testifying or you can return to your seat. Witnesses who are unable to physically attend the hearing today can submit written comments via the Legislative portal. Please note that any written testimony transmitted is considered public comment and may be read into the record. So once again we are so grateful to have each and every one of you joining us here today.

And now I'd like to make some opening comments before passing it along to my Co-Chair for this hearing, Assemblymember Bauer-Kahan. Some days it feels like we're living in an episode of Star Trek, doesn't it? We carry the collective knowledge of all the world's civilizations in our pocket.

We can instantaneously communicate with people on the other side of the world. Medical advances regularly treat diseases that only a few years ago would have been catastrophic. Each new generation has been forced to grapple with the rapid pace of technological innovation so that our children and their children can enjoy a brighter future.

The California State University is truly one of the crown jewels of our state. In a place where our students go to develop skills to meet the moment. Universities have always been great laboratories of our society. The places where we go better understand ourselves and our world. The places where we interact with new ideas and yes, new technologies.

Artificial intelligence is one of those new technologies. As we've seen in just a short time, AA has become a powerful tool that is widely used across sectors. In February of this year, the Chancellor's Office announced the decision to become the nation's first and largest AI empowered University system. Through a partnership with OpenAI.

The CSU has rolled out ChatGPT EDU and as I understand it, over 40 over 93,000 ChatGPT EDU accounts have been activated system wide. That is more than twice the number originally projected for this launch period. Only one element of a much larger AI integration that we'll be exploring this afternoon. I'm eager to learn more about this integration and to learn more about the development of this initiative and also the rollout of this initiative and the future short term and midterm planning.

I also want to better understand the training that students, faculty and staff are receiving to best utilize these new technologies, especially technology that has been widely reported, can sometimes very confidently provide very incorrect information to our users.

There are also questions of ethical use and how this technology will impact our higher educational institutions in the classroom, but also organizationally. Thankfully, we also have a panel of leaders from the CSU that have been involved in the development of this initiative. Dr. Clark, Dr. Kennedy, Dr. Hirsh, appreciate you coming today and providing your testimony today.

I'm also looking forward to hearing the experiences from our critical stakeholders within the system. I understand the level of input and consultation that you have all been able to to offer, as well as your experiences with this technology that is being rolled out. The leaders in our second panel would also provide the critical insight.

Dr. Boyd, Dr. Bernal, President Hutchinson, and Vice President Coriam, thank you so much also for taking the time to be here today. And while I'm saying thank yous, also like to thank our Chair, Rebecca Bauer-Kahan...your Committee, the Privacy Committee, and your staff for the incredible oversight that you have all been taking regarding AI. And I'm very happy that we'll be working together on this. I'd like to pass it over to the Chair of the Privacy Committee, Assemblymember Bauer-Kahan, for opening remarks.

Thank you, Chair. And I want to start by thanking you for your leadership in initiating this oversight hearing and inviting me as Chair, but also Committee Members to join you for this really critically important conversation. And I want to thank Committee staff who I know work very hard. And this is supposed to be when things are a little quieter for them. So I appreciate them putting this hearing on right now.

During this time, the sergeants, other Capitol staff, and the panelists who are here, I'm so happy that we have today a gathering of leadership, administration, faculty, staff, students, the whole ecosystem to really talk about how this rollout is happening and how they're engaging with the technology.

Over the last two years, our Committee, the Privacy and Consumer Protection Committee, has held or participated in hearings on AI ranging from issues regarding the arts, workforce, healthcare, and more. And the thing that we have, the common thread throughout all of these hearings is a question of how do we integrate AI to harness its potential, but manage the disruptions. And I think that is really the foundation of how we have tried to approach this, which is AI has so much potential. It is such an exciting innovation.

California is leading the way in being the home of so much of that innovation, and yet we know that with it comes risk. And so we have to see that innovation through the lens of how it can help us while protecting ourselves from the harms that can potentially appear from it. Education is one of the most consequential contexts for AI. One of the reasons that California leads the way in so many ways through technology is our incredible education institutions. It is the talent that the CSUs, the UCs, and our private institutions create and put out into the world to make the world change through things like artificial intelligence.

And we need to meet our students in the moment we are in, which means helping them to prepare for the future that they are facing, which is one that will be very much AI centric. I want to thank the University for really leaning in to the moment. AI literacy is essential to making sure today's graduates are using the tools in a responsible way. These tools, such as ChatGPT, which was mentioned by the Chair, can help students with complex issues in really incredibly personalized ways.

It can unlock research and provide faculty with support that allows them to focus on their students and the academic outcomes we want them to be focused on. So I really want to thank the CSU for that and for everything they're doing to be here today. But also focus so much on ensuring that we have successful students that we are creating in the great State of California. But I would be remiss if I didn't talk about the important things that we need to remember as we face this AI future. We need to do so in a way that is inclusive.

We know that like all tech tools, there is a way where many people can be left behind. I saw data last week that showed that adoption amongst women pales in comparison to adoption amongst men. That is concerning for me because as the workforce becomes more AI centric, those women can very quickly be left behind.

Which is why this project can be so central to creating equity, but only if we do it properly. In addition, we know tools can go wrong. They can have bias baked in with biased data. And so we need to make sure tools we're deploying are responsible and, and they create the society we want to see. I want to say that I'm happy the workers are here today because I think one of the ways to best create the future we want in an AI world is to put the people at the table who are most impacted by the artificial intelligence.

And in this context, I truly believe that is the CSU's workers, the CSU faculty, and the CSU students. And so they need to be at the table driving how we integrate this in a safe and responsible manner. Privacy - AI has created the next gold rush. And that gold rush is for our data. And so we need to make sure that as we create these partnerships, we understand the trust that Californians put in us when they share their data with us. They don't feel like they're sharing their data with a private company when they share it with the State of California.

And so as we take that data, we need to be responsible with it. And so I look forward to conversations today about the contracts and how they will protect the privacy of students and anyone who engages with these tools. Lastly, I would be remiss because it is something that is so deeply, I'm passionate about. But we are seeing a way in which these tools are becoming the next generation of social media. We are seeing people engage with them in addictive fashions. And just this week, the news was broken that a young man took his life at the coaching of ChatGPT.

And so, as we take people's children out of their homes, maybe for the first time, and we become responsible for them in an academic setting, we, the State of California and the CSU, have got to make sure the tools they are putting put into their hands in our institutions are safe, that they are not sycophantic, that they are not addictive, that they are engaging with them in a way that will protect their mental health and allow them to unlock the educational potential. And so all of these things need to be central to what is happening in our conversations around this. So I'm excited to hear about this today, and I want to thank the Chair for starting this conversation.

Thank you so much, Madam Chair Bauer-Kahan, for your opening, powerful remarks. And also grateful to the Committee staff and everyone involved with today's hearing for organizing both of our committees in this collaboration here today. I'm also grateful to be joined by a number of our colleagues as well. Would any of our colleagues like to make any opening remarks? Thank you so much and welcome. Now we'd like to introduce our first panel, CSU's Artificial Intelligence Empowered Initiative.

Our speakers for the first panel are Dr. Ed Clark, the CSU's Chief Information Officer, Dr. Leslie Kennedy, CSU's Assistant Vice Chancellor of Academic Technology Services, and finally, Dr. Sandra Hirsh, Associate Dean for Academics and College of Information Data Society at San Jose State University. Each speaker will have six to eight minutes to provide testimony. Thank you so much for all being here today. You may begin when you're ready. Welcome.

Thank you. Chair Fong and Chair Bauer-Kahan and Members of the Committee, thank you for the opportunity to testify today. My name is Ed Clark and I serve as a Chief Information Officer for the California State University System, the nation's largest and most diverse four year public university system. At CSU, our mission is rooted in equity, access and student success. That mission is at the heart of how we are approaching artificial intelligence. We are not adopting AI because it is fashionable, but because it can directly serve our students, faculty and the State of California.

And we've gained many willing partners in California and beyond thanks to the CSU's uniqueness and unmatched potential to educate new generations of leaders in every corner of our state who are passionate, equipped and ready to contribute their limitless talents to California's increasingly AI driven workforce. Our work in artificial intelligence is guided by strong shared governance.

It began with a statewide Academic Senate resolution in 2019 and a second one in 2023 recommending the establishment of a System Wide Generative AI Committee to plan for the impact of AI in higher education. This Committee was formed in 2024 and as an official Committee of the CSU, comprised members from stakeholder groups from across the system, including three faculty members along with staff and students, all appointed by the respective governance bodies. The Generative AI Committee also formed subcommittees to explore issues ranging from teaching and learning to security and privacy.

These subcommittees consisted of even more faculty, staff, and student appointees from across the system. The Generative AI Committee and its subcommittees met regularly throughout the spring semester of 2024 and this work culminated in a widely disseminated final report and recommendations.

Perhaps one of the most important recommendations in the report was, and I quote, ensure inclusivity and equitable access. The CSU should promote inclusive and equitable access to Gen AI technologies and ensure that all CSU faculty, staff and students have access to Gen AI tools and training necessary to leverage them for teaching, learning, research and work.

This final report of the CSU Generative AI Committee was the foundation of the System Wide Generative AI Strategic Plan. The plan presents a broad strategy grounded in values of ethical and equitable AI use, prioritizing governance, training, inclusive access, protection of intellectual property, data stewardship, and ongoing learning to enable responsible and transformative adoption of Generative AI across the system. This governance matters because AI Adoption is happening rapidly and often in fragmented ways.

Prior to the creation of our AI strategy, thousands of CSU students, faculty and staff were already signing up for individual AI tool subscriptions, yet many of our community members could not afford such subscriptions. Some of our universities were exploring their own licensing plans with various AI vendors, yet not all of our universities could afford such investments.

Indeed, universities across the United States are facing the challenges of uneven access and inconsistent protections amid growing workforce expectations. We knew that the CSU needed a coordinated, responsible approach that would expand opportunity rather than deepen divides. Let me outline our CSU AI strategy with some of our guiding principles.

First, Equitable Access. Responding to the Generative AI Committee's recommendation for Equitable Access, much as we do for other technology platforms, we negotiated system wide contracts so that all 23 campuses and our 460,000 students and 50,000 faculty and staff could have access to the same AI tools with strong privacy protections.

This ensures that students at resource challenged campuses enjoy the same opportunities as those at larger campuses. By providing access for all, including low income students and underserved populations, we are preparing our students for the workforce of the future. They will have the familiarity and the capability to leverage AI in their careers post graduation.

Two. Faculty are at the center of this effort. We want to emphasize that this is voluntary. Many faculty have been involved, from the initial Senate resolutions to the system wide committees to CSU's current efforts. Recently, more than 700 CSU faculty applied for Teaching Innovation grants to integrate AI in ways that support critical thinking, ethical use, and creativity.

My colleague will share more details about this in a moment. Three. To meet the CSU Generative AI Committee's request for universal training and tools, we launched the CSU AI Comments to share best practices, training and resources across all campuses. This site and its contents are freely available to anyone in the CSU community.

In addition to AI training courses created by our own CSU faculty, we also host training programs provided by our industry partners including Google, IBM, Microsoft, and OpenAI. And the list is growing. The AI Commons will ensure that the full CSU community will have access to, learn and grow AI skills to advance their careers.

Finally, through the AI Workforce Acceleration Board, we're working hand in hand with state leaders and industry. This board is helping the system ensure that our graduates are prepared for California's evolving workforce needs and that the CSU is aligned with statewide economic development strategies.

The Board includes faculty, staff, campus representatives, two officials from the Governor's Office, and industry partners. Those industry partners have committed to building a variety of internship programs and opportunities for our students, along with a host of other benefits for our CSU community. These include the AWS Summer AI Innovation Event, which included students from every campus in the CSU building working AI solutions to solve real world problems.

We're also piloting the concept of Project Based Paid Micro Internships with many of our industry partners beginning this fall semester. As you can see, this strategy is already producing results across our system. My colleague Dr. Leslie Kennedy will now share some of the outcomes we're seeing and how CSU is positioning itself in the State of California for excellence in student outcomes and workforce preparation.

Thank you, Dr. Clark, and thank you again to the Committee for inviting us. I am Dr. Leslie Kennedy and I serve as the Associate Vice Chancellor for Academic Technology Services. Previously, I have been a tenured faculty in the Humanities. Before you hear from our campuses, I'd like to share what this looks like on the ground across the CSU. As Ed noted, in 2019, anticipating the transformative impact of AI on employment, higher education and student success, the statewide Academic Senate created a resolution requesting the Chancellor's Office establish a working group to examine the implications of these changes and propose actionable recommendations to address emerging challenges.

Following the Academic Senate suggestion in 24 we formed the System Wide CSU Generative AI Committee and subsequently developed guidelines, governance structures and collaborative strategies to ensure the ethical and effective integration of generative AI across the campuses, prioritizing human involvement, risk mitigation and CSU values.

A key recommendation was to reduce the digital divide by providing inclusive access to a protected AI environment for all CSU users. In response, the CSU secured a system wide license for an AI tool granting students, faculty and staff access without the financial barrier of individual paid subscriptions.

Tailored specifically for the CSU, this tool delivers advanced AI capabilities alongside enterprise level privacy, data protection and security features such as single sign on. Interactions within the CSU, AI tools are treated with the same confidentiality as other CSU platforms such as email. CSU does not monitor these communications.

We have put policies in place that are designed to support faculty and students to protect their privacy and academic freedom. The CSU AI license further safeguards data, guaranteeing that no CSU data prompts or responses will be used to train large language models or third party systems. These safeguards are critical in ensuring that our faculty and students intellectual property is kept as theirs and not used by a vendor's large language model.

In spring 24, the Academic Senate of the CSU approved a resolution commending the Chancellor's Office for funding and offering system wide professional development opportunities that allow faculty to engage with and explore AI tools for teaching and learning.

The resolution also encouraged expanding these opportunities to deepen faculty understanding of rapidly evolving AI Technologies, exploring strategies to uphold academic integrity, the strengths and weaknesses of integrating generative AI into pedagogy, and the potential barriers and benefits for CSU student success in the 21st century.

Since spring 24, multiple online introduction to teaching and learning with AI training opportunities have been made available at no cost to faculty, staff, and students. More than 4,300 faculty Members have already voluntarily participated in these courses, facilitated by the chancellor's office. The 2024 academic Senate resolution specifically praised the quality of this training and encouraged wider flexibility.

Faculty Participation Several campuses, including CSU, Monterey Bay, Fresno, and San Diego, developed the original training programs, which were so effective that they were scaled system wide in 24.

Spring of 24 the scope of AI training has since expanded to how users can explore key principles such as ethical and responsible use, critical thinking, and prompt design, recognizing the limitations and challenges of AI systems and and exploring diverse applications of generative AI.

In June 25, the Chancellor's office awarded $3 million to 63 faculty led projects selected from over 400 submissions involving 700 faculty Members. System wide. The challenge invited faculty to design innovative instructional strategies that leverage AI tools to integrate AI fluency into curricula, strengthen critical thinking, and promote ethical use of AI.

The grant proposals were evaluated for their system wide impact, particularly their potential to serve diverse learners across disciplines. The grant program was guided by three core principles AI Literacy, critical Engagement, and empowered Ethics.

Every CSU campus is represented among the awardees, and the funded projects include disciplines spanning engineering, history, ethnic studies, health sciences, teacher preparation, scholarly writing, and journalism. Many projects have adopted interdisciplinary approaches or emphasize faculty development, presenting instructors with tools for instructional transformations in an AI enabled environment.

Most funded proposals also include extensive research components focused on identifying best use cases for transformative student learning experiences. These findings will be shared across the campuses to maximize system wide impact, enabling faculty to build one another's instructional strategies to prepare CSU students for success.

With this initiative, the CSU has positioned itself as a national leader in supporting faculty research and advancing curriculum transformation through AI. The CSU's AI strategy matters because artificial intelligence is reshaping every sector of society, the economy, and higher education.

By moving early and intentionally, the CSU is ensuring that students are prepared for the future of work, equipping graduates with AI literacy, critical thinking, and ethical frameworks ensuring they enter the workforce with skills employers urgently need strengthening California's economy, equity, and access remain central.

By providing system wide licenses, faculty development and student training at no cost, the CSU is reducing the digital divide and ensuring that AI opportunities are not limited to those who can afford them. Faculty can lead innovation.

Supporting faculty with grants, voluntary training, and collaborative governance structures empowers them to explore and integrate AI into teaching, research and service in a secure environment. Partnerships with the State of California and industry will amplify impact and trust and responsibility will guide adoption.

In conclusion, by introducing AI into the fabric of teaching, learning and workforce development, the CSU is demonstrating how a public University system can expand opportunity, inspire innovation, and prepare the next generation to lead a rapidly changing world.

Our current progress is undeniable and we are beyond excited for all that this initiative will achieve for our students and their families and communities and for our Golden State. We are excited to be working with faculty, staff, students and our campuses on this important effort.

Thank you for your time and attention and we look forward to your questions.

Thank you. Welcome.

Thank you very much. Thank you. Chair Fong, Chair Bauer Cahan, and Members of the committees for holding this important hearing on Artificial Intelligence in higher education.

I'm Dr. Sandy Hirsch and I most recently served as Associate Dean of Academics in the College of Information, Data and Society at San Jose State University and I really appreciate the opportunity to share how the CSU and San Jose State University are proactively addressing this critical topic and preparing our students for an AI integrated future.

First, at sdsu, we are actively evolving the curriculum to incorporate AI into courses and develop new academic programs to ensure that students are well prepared to meet changing workforce needs. For example, we already offer 13 AI focused programs at the Master's and undergraduate levels such as an Ms. In AI and an Ms.

In Applied Data Intelligence, as well as five minors and certificates across the University. We are also offering more than 60 AI focused courses and we have more courses and programs in the pipeline.

Current courses range from Storytelling in the Age of Intelligent Machines and Mass Communications to Ethics of AI in the School of Information to Introduction to AI and Software Engineering. These courses are thought of and designed by faculty.

One new program I particularly want to highlight is the new Human Centered AI Certificate that we have launched this fall 2025. This undergraduate certificate is open to any major and we were pleased to see that students from a wide range of disciplines and colleges of staff signed up for this new program.

This certificate offers students a fundamental overview of AI technologies focusing on practical applications and equipping them with the skills to critically assess AI societal implications. We are also working on expanding this into a more in depth interdisciplinary minor in human centered AI, which we plan to launch in fall 2026.

In addition to these programs, we also continue to adopt existing courses to incorporate AI and enhance student learning.

For example, the First Year Writing Program at San Jose State has integrated digital and creative literacy tools through Adobe, including AI tools such as Firefly for quick video editing into the writing program so that students can apply writing to concrete examples such as film scripts, biographies or even new product ideas.

Our Career center under the direction of the Division of Student affairs is also offering a new two day jump starting Digital Literacy program. This co curricular educational event is being held in collaboration with the First Year Writing Program, the Dr. Martin Luther King Jr. Library, the San Jose State Writing center and Adobe Student Ambassadors.

This center is also working with IBM Skills Build to plan a hybrid hackathon for SYesU in October 2025 with a focus on responsible AI aligning with the newly launched Human Centered AI Certificate, our AI Literacy course, and the AI Innovation Hub and partners.

Additionally, some of the AI Educational Innovations Challenge grants that were awarded to faculty at San Jose State will further help them consider how to integrate AI into critical courses.

For example, one of these projects specifically focuses on redesigning two foundational GE writing courses, English 1, B and 2, to embed AI literacy, critical thinking and ethical authorship through assignments that include AI, human co writing and multimodal storytelling. It will also develop and offer faculty workshop series Teaching AI for Writing instructors.

In addition to building and evolving the AI future focused curriculum and co curricular experiences, we are also building administrative capacity and collaboration across the University to ensure alignment on AI related initiatives and we're providing opportunities for faculty and students to gain AI skills.

One example is a new cross campus collaboration on AI pedagogy and curricular inNovations called Nova, which stands for Network for Optimizing AI.

And this group includes representatives from across campus in student and faculty supporting roles ranging from the Career center to the library to the Center for Faculty Excellence in Teaching Innovation to college Associate Deans and it's to make sure that we're aligning our efforts across the campus. Faculty and student representatives will be added this semester.

San Jose State is also actively preparing students to be AI literate and to use AI responsibly. This summer at orientation, we launched a new AI Literacy program in Canvas called AI Literacy Essentials for Spartans for all incoming San Jose State undergraduate and graduate students.

This course provides visibility into the AI tools available to San Jose State students and addresses how to evaluate AI generated content for accuracy, bias and appropriate use. Additionally, we are working with faculty on how they can consider integrating AI into their instructional practices. For example, we have just launched a new AI Faculty Fellows Program.

In this year long program, a faculty representative from each college serves as a strategic liaison and thought leader for AI enhanced teaching and learning within their own discipline.

Through this initiative, AI Faculty Fellows work within their own colleges to help uncover innovative practices, support peer learning, and contribute to a growing culture of responsible, ethical and transformative use of generative AI in teaching and learning. The faculty involved will also share their learnings and best practices across our University campus.

This fall we will also be extending this program to create a parallel Student AI Ambassador program. One student from each college will serve as a peer educator, Community Liaison and creative collaborator, directly supporting the integration of AI from a student perspective. Working with the AI Faculty Fellows Committee Members.

I hope this has conveyed some of the work that's being done by faculty, students and campus leaders at San Jose State on AI.

From Student Orientation to AI Education Innovation Challenge Grants, Our goal has been to work with faculty and campus partners to ensure that students graduate workforce ready while emphasizing critical thinking and the ethical use of AI. I believe we're on track to meet those of the objectives. Thank you.

Thank you so much. And thank you so much to each of our panelists for your testimony here today. Really appreciate it and thank you for your sharing your insights. Open up with a couple questions and I'll pass it to my colleagues for additional questions. A number of you talked about the ethical use of AI.

We know that AI models have been shown to perpetuate existing biases at times raising concerns about potential discrimination in areas like hiring support services and academic outcomes. How are users being trained to recognize potential biases? And secondly, were students, staff and faculty required to receive training before accessing chatgpt. Edu?

This technology, as we've acknowledged, is so new and at this time each of our campuses is individually creating their own individual decisions about required training. Certainly we're open to additional system wide requirements as this technology evolves, but some of our campuses, as you just heard, have these requirements emerging for entering students, for example.

So I'll just add that like at San Jose State I mentioned the new orientation, students are now getting this AI literacy training as part of all of our new incoming students are. And this fall fall we're also rolling out the same kind of training, AI literacy training for current students and also for graduate students as well.

And also for our faculty. As you heard, there's a lot of different initiatives for providing them with opportunities to Learn and to as they're interested in acquiring and learning more about AI literacy.

Thank you so much for that context. And just to follow up on that, you mentioned AI student ambassadors, AI literacy training in terms of the San Jose State model. Is that something the chancellor's office can scale up system wide or best practices?

Yes. So most of our training that we're providing currently includes ethical and responsible use training or discussions or information including the biases issues that you mentioned. We scaled that up from San Diego State's activities that were happening that they created at their campus and we continually revise these courses.

So we shall be doing that again for this round for this academic year. And we'll take particular note of those areas with regards to what San Jose State's doing.

And building on that, we mentioned the CSU AI Commons. We host all these trainings in an area that anyone in the CSU can access for free. It's open to all of our CSU community Members.

Great. Thank you so much. Also like to note that Ruthie Barku is here from OpenAI is here in the audience and here to provide any technical assistance. Thank you. Open it for my colleagues, Chair Bauer, Karin or I'm sorry. And then we'll follow.

I have a lot of questions, but I do I have one follow up to Chair Fong's question.

So before I turn it over to Patel, what I just heard you say that as it relates to what I think the Legislature believes is a critically important issue, which is addressing bias, there is a patchwork of policies across your system and that the CSU itself is not providing guidance to ensure there is coherent policies at all of our campuses.

So I want to give you an opportunity to correct that because that's what I heard, but maybe I misheard.

I think it would be accurate to say that, you know, this AI strategy was approved by our board in January. We've started rolling out the tool, the ChatGPT tool, about midway through the semester. So we've only had a couple of months to even get to this.

Now over 100,000 activated users building the training and providing the support in the CSU AI Commons is how we're approaching this. So we that training does include, it does address things like bias, ethical thinking, critical thinking. And so that is where we are at this moment with this very new technology.

And on our AI Commons we have a space for each one of our roles, a role in the system, students, faculty, staff, and we have recommendations for ethical and responsible use in those 34 areas, including alumni.

Okay. So I Guess my concern as I'm listening to this is, you know, you rolled it out, you have more users than I think any of us, including you, would have expected, which is good news in a lot of ways. But I also want to make sure that we are being responsible in this.

And so, you know, sort of. There have been some in the AI community that believe the way we deal with safeguarding is to put it out, see what bad happens, and then address it. I think we as government have much more responsibility than that and want to make sure that is not what is happening at the CSU.

It sounds like you have certain campuses that perhaps are leading the way, but I would hope that to the extent that that is happening, it is immediately being rolled out across the entire system, rather than allowing each to do their own thing.

I think the expertise perhaps is at the system wide level and that expertise should be making its way to each of the campuses. With that, I'll turn it over to Salivary Patel.

Thank you. Dr. Patel, thank you.

Thank you for being here and presenting the information that you've done today. I want to commend you for taking a very proactive approach and being very inclusive of all the people of the CSU community. That, that is great to hear. A couple of questions, a little more technical maybe.

I just want to make sure I heard correctly, Dr. Kennedy, when you said that the large language models are sort of internal, they don't allow for anyone to take any information out of it to make our students information vulnerable. Are these closed AI systems that you've now licensed from OpenAI?

Correct? It's in our license agreement.

And how is it using the inputs and outputs to develop its own, to evolve in its own internal system? And at what point do you then open it up to new inputs from external sources? How often does that process happen?

I'm not familiar with that algorithm or that dynamic, are you?

So we definitely working with it. And actually we have an expert from OpenAI that might be better positioned to answer this question. But we definitely get upgrades. Whenever OpenAI upgrades their model, we get the same upgrade, but our own data and interactions are all held in our own private space.

So that's why they can't be used by OpenAI or any other third parties to train models.

Would we like to have the technical expert join us? Is Ruthie here?

Yes, she's here.

Please come on up. Thank you. Right here.

Okay.

Do you have anything you would like to add to that?

I agree with Dr. Sorry, Madam Chair.

Please.

Thank you for the question. Representative Patel, I agree with Dr. Clark's comment. So we do not have access to the data. The IP and the data is ultimately owned by the student. I defer to, you know how this is administrative by the CSU system as well. And then we also do not train on the data.

Okay, thank you. So if the data is self contained, is there any assumption that each individual user has privacy of their own inputs? So the University system is not looking at, has no awareness of what inputs are being put into the system?

Correct?

That's correct.

If a student is to encounter something that they, or a user, is to encounter something that they consider biased, is there a flagging system for that?

There are ways that you can report, you know, that you're unhappy with an interaction and actually it's partly baked into the tool. But certainly we're expecting our community to let us know about that, those things when they arise.

And is there a system like is there a checkbox on a website that says hey, flag what you got from OpenAI?

The thing is when we're in ChatGPT, OpenAI, our version, there's a space where one can report those kinds of.

That's great, it's good to hear because I think our young people today have a great sense of awareness and they want to be proactive in how they engage with AI. They are natives. They are AI natives, so to speak. And so they're very open with pointing out bias.

I've got three young people, a 20 year old, a 17 year old and a 15 year old and they're very clear with what they see as fake information. They will share it back with me.

Another question I have is that is there a plan or maybe it's already ongoing for working with our local high schools in the region for helping them develop best practices?

Maybe if you have high school students that are dual enrolled or perhaps you have AP courses that are intended to be college level courses, are we able to provide any of that information to our local school jurisdiction so that they're also helping prepare students for what they're going to encounter on a college campus?

We're very aware that we teach over 50% of the teachers or prepare the 50% of the teachers in the state.

So we are focusing on those pre service teachers and providing several of our grants to the, to those areas so that they can strategize, create different ways of preparing students in their coursework in the future or their teaching environments.

And we do have a ways to go though, to go to actually I think you're saying be part of the K12 system.

Providing best practices because those high school students who may be dual enrolled in a college class or in an AP class where, where they believe they are taking a college level class, AP classes are supposed to give that environment of that same academic level of achievement.

And if they're enrolled in our system then they can get access to any of our training and resources, even the closed version of ChatGPT as well. But we will take that back as an advisement and see what we can do moving forward.

Just support and guidance for those high school campuses.

Yeah, we were thinking more of the teachers preparing them for that as well. So. Yeah.

And that trickle sometimes doesn't permeate.

Yeah, yeah.

So thank you.

Thank you.

Thank you so much, Dr. Patel. Madam Chair Bauer-Kahan.

Thank you. I'm going to, first, I want to just start with my own comment but then I have. Assembly Member Pellerin had to go to an important meeting so she asked me to ask her questions, if you don't mind.

But before I do, I wanted to say that I did read the report, the Joint Committee on Education Policy and Finance report on the project and it included some highlighted classes that you were teaching or projects and if they're classes or projects on AI ethics.

For anyone that hasn't seen that, it was actually very impressive and I think that this is really exciting. I will say one of, I believe it was last week, there was the news broke around meta's training document for their AI model. It had been approved by their chief ethicist and it included the ability to ask a child.

The training document said that it was acceptable to engage a child in conversations that are romantic or sensual. The chief ethicist at META approved this. It also said that it was unacceptable to describe a child under 13 in terms that indicate they are sexually desirable.

But a 13 year old, apparently it's okay for the models to describe as desirable. To me what this actually said to me was that we have not done enough to dive into what are the ethics of training these models such that we have people in the roles to read that and say clearly this is not ethical training. We don't want children to be treated this way in society.

And so I think that the University system diving into what are AI ethics in a way only universities frankly can is actually a really exciting thing because I think it will allow us to engage in conversations that produce the next generation of AI ethicists that will lead in responsible ways.

And so I really wanted to highlight that because I think that's an amazingly exciting thing that the universities are doing. So with that, I want to turn it over to Ms. Pellerin's question. So the first one was on environmental impact.

She wanted to know if the University, if the CSU system had looked at the environmental impact in terms of energy and water usage of your immense number of students that are going to be using these models, and if so, what are you doing about it?

Yes, we're contracting. We want to track all of the sort of the impacts of what we're doing, both positive and negative. So we have an actual metrics initiative where we're collecting everything from adoption to potential environmental impacts and how we would measure those things.

And so that metrics effort is not going to be completed for another month, but it's one of these things that we're trying to track.

That's great. And I will say now, speaking of myself, just so we're clear, I think that again, this is something that our universities can really help us understand how to navigate in the future.

And hopefully we can take your own personal experience and turn it into learning such that the private sector and us can be more responsible in the environmental impact that we have as we continue our AI revolution. Okay, her second question, we know that the use of AI to generate image based sexual abuse is becoming increasingly widespread.

Is there easily accessible and very active blogs and messaging forums about how to bypass the safeguards of the models to generate sexual content? And frankly, I'll add this, I think there are some models that don't have those safeguards. We have someone sitting here. I think ChatGPT does have those safeguards, but not all models do.

What additional safeguards have you developed for this version to make sure that your students are not getting around the protections that are built into ChatGPT.

These sort of malicious prompt injections, those kinds of things that people might use, would violate the CSU's acceptable use policies. We have acceptable use policies that prevent you from committing crimes on the Internet, for example. And these same acceptable use policies apply here, to how you would use this tool.

And so then she wants to know, given that it would be a violation of your acceptable use policy, what will the disciplinary process be? Will the first offense result in expulsion? Will the disciplinary process for misuse, is it clearly stated? Do students know what happens if they misuse these tools?

Academic discipline is something one of my colleagues here could probably answer better yeah.

And this it will and I'll take defer to you in a second. But every campus has its student conduct policy so it would depend on the campus where that activity happens as to how they would be disciplined. And so I'm not sure if you.

Yeah, I don't have anything more to add. I mean that would. That we do have academic discipline processes on campus and so it would be handled along those lines as merited.

Okay. Well I hope that as you are onboarding students, I mean non consensual notification which is something that is happening and is completely upsetting and destructive to people and their mental health. I hope that students are being, this is expressly addressed and they understand the consequences and the consequences are real.

I would hope that it was consistent across the system. For some reason I'm surprised maybe I don't sit on higher ed. I'm surprised that there's not more consistency across the system. Go on.

We need more consistency. That's a good one.

I was just going to say that's part of the impetus for us to create that nova that I was talking about then that cross campus collaboration at San Jose State is to bring together the various leaders who touch students in different ways whether it's through writing or through the student services support wellness looking across so that we can make sure that we treat students and faculty, like deal with it in a holistic way and to flag any kinds of concerns that are emerging so that we can address them in a more systematic basis.

Yeah. And obviously the model can't anticipate everything that's going to happen and it is upon students to use all technology responsibly. So we need to see that. Her next question is we've seen increased reports of people using ChatGPT as therapists, although I think that's a generous term.

They're not actually therapists but, and falling in love with AI partners. What safeguards have you developed to protect students from this?

The security and safety of our students is paramount for us in the CSU. We are always trying to look at how any technology might be again used or misused. Right? And it's a very new topic for us and we're still developing our policies and thoughts around this.

Certainly we, in the same way we wouldn't want our students to use social media to get coaching on doing harmful activities. We wouldn't want that to happen on ChatGPT either. But it's a very serious issue for us to teach our students about because much of what they access is actually outside of the CSU's control.

Right? But you've put this in their hands, which I think is why we are.

This closed space. Yeah, we put a closed one in our space.

Yes. And so I will now do deviate from my colleague's question and say, and I had this conversation earlier today with OpenAI, so it is widely noted, and I think they've been transparent about this, that GPT 4.0 was friendlier, warmer, more sycophantic than GPT 5.

And I will say one of my children remarked to me, actually, a noted difference between the two models. That was a choice that OpenAI made. You know, there's also been reports that people didn't like it when it wasn't as warm and nice, and so they've dialed it slightly back towards 4.0.

So those are choices that the models are making. Is there anything in your contract that specifically addresses sycophancy?

Not at this time. We don't address that at this time. But certainly, as I.

Put that on your list, because I do think that if we're putting this in the hands of students for academic purposes, the tool does not need to be sycophantic. Frankly, I don't know that it needs to remember personal characteristics of the students, which is one of the things that people are starting to believe is causing some of the relational addiction.

Those are things that do not have to be in the model and are choices that are made on models. And we have the right, I believe, in these contracts to determine those things. And we should be. And safety should be paramount, and we should be making the safest decisions for our students. And none of that would deplete the ability for the academic outcomes that I think everybody at this table and this dais wants.

And so I would ask the CSUs to look at that directly and come back to us in short order on how you're going to address that. Her next question is, will student support services be immediately notified if a student is in the EDU system, in your closed system, talking about suicide?

So my colleague here addressed the fact that we treat the CSU interactions are totally private. So we treat it just like email. We don't actually see those interactions, just like we don't see your email.

And so there's not really a way for us to just say, hey, we saw that you're doing this or that or the other until. Because it would be violating people's privacy otherwise.

I will note if someone walked into your health system and told one of your physicians they were thinking about suicide that would not be covered by doctor/patient privilege if it was a realistic concern of theirs. They are a mandated reporter and so the privacy would not extend to discussions of suicide in that instance.

And I think you should reconsider your privacy in this instance as well. So one of her questions here, which I think is an interesting one, is has the University system considered liability as it relates to a student that uses these systems to harm themselves or to create non consensual pornography?

This is a question that would be better served by our legal team that worked on the contract. I'm not aware of that level of detail.

As the individuals that fund you and oversee your budget, liability is obviously a piece of that and it'd be important to us that, that you are protecting yourselves from potential liability in these ways. So again, would appreciate you following up on that and your council can do so. This one, again, Ms. Pellerin's question.

How do you respond to the numerous studies that have come out recently that show that ChatGPT used by students leads them to get to getting, quote, lazier over time? Specifically the MIT study that came out recently that found that ChatGPT users had the lowest brain engagement and I quote, can consistently underperformed at neutral linguistic and behavioral levels.

So we're aware of the MIT study and other studies that are also either in support or potentially showing other strategies around that.

So again, we are encouraging our faculty and our students to utilize the tool if they in order to prepare either for the workforce or to be able to still use our critical thinking skills, for example, in their prompt exercises and other ways that require critical thinking.

So we're hoping that it's not encouraging lazy thinking, but that the students are actually learning different strategies around how to use AI to support their in learning and teaching environment or the learning environment.

I appreciate that and I'll say, speaking for myself now, I actually really appreciated the many, both in the report I cited earlier and in the final report you put out the discussion on literacy that comes with this.

I think it's really important we know and it's no secret that these tools aren't always 100% accurate and that it is really important for students to be able to check the primary sources to make sure they're getting the right information. I'm a former practicing attorney.

I have an old colleague who is constantly testing ChatGPT on its legal and always putting out these funny stories of all the wrong advice it gives. And that's not, you know, I mean, these tools are not perfect and so I do think the literacy piece is really critical.

Yeah. And we funded an activity with libraries. Our libraries have entered into a AI literacy research project around that. And so have our Centers for Teaching and Learning, several of them as well. Around AI literacy.

Yeah. Okay, so then. And I don't. I'm not aware of this, but this may be a question for you. In July, a federal judge ordered the ChatGPT preserve nearly every exchange its users have had with the chatbot. Is that something you're aware of? Okay, you're both aware of that. So how are you keeping this data private while also complying with that court order?

Because, because we do have our private instance. We weren't subject to that subpoena.

Okay. So that's separate. Okay. And then, as was mentioned by our colleague, ChatGPT is a closed source model. Does the CSU know the model weights embedded in the code? Like how much, you don't?

No, we dont.

You already answered that question. And I will say, as someone who is a little bit of an intellectual property nerd, I do appreciate the effort that is going into protecting the intellectual property at your universities.

And I think one of the things that I think is really exciting about this, frankly, is we don't pay enough attention here in the state to the way we can impact policy through our purchasing power. It's real.

And, you know, we are, I would guess, one of the largest purchasers in the state of many, many things, the state itself. And so, you know, making those kind of requirements on a contract, I think changes the game and it empowers others to do the same. And so I think that's really powerful.

And I think you've done a couple things that we've heard today in the contract that are novel and interesting and I think will hopefully be a way for others to follow in changing the way we purchase these tools, especially as it relates to government, specifically, because I do think we have a different obligation with data than private sector purchasers.

So I think, frankly, she asked a lot of my questions, which was great. One thing that I think, I guess the privacy thing, I think you've been pretty clear about it. It's all private, it's all separate. It sounds like they don't have access to any of the student data. That was my understanding. Right?

That was what was said. And so, and you're not monitoring any of the activity of staff or faculty. What if there is an accusation of plagiarism or some misuse of the tool? Can you access it at that point? No. It's always private. Fascinating. Okay.

And then in other moments where the government is engaged in AI in this way, we've used a sandbox approach. And where it's controlled and limited, here you sort of just put it out into the CSU system with nearly half a million students opting in.

I guess, and I'm hearing a little bit of, you know, your presentations are great. The minute we got into questions, it felt like maybe enough thought hadn't gone into some of these policies, frankly, that I know you're going to follow up on. Why did you choose not to pilot it? Why did you choose to go all in at once?

Well, I'll try and invite my colleagues to also jump in. You know, the CSU was not even close to being the first to adopt these tools. You see, large universities, Michigan, Harvard, Florida, Arizona State have adopted these tools.

And what we were most concerned about is that we would be causing a digital divide with our CSU access institutions versus these very large, better funded R1s. And so our thought was like our students should not be left behind. They cannot be left behind as this economy and this workforce evolves. And that's why we decided to move forward much as these other universities did.

But the campuses were able to implement in their own timelines. So some were more, were able to move forward faster than others. Others took a more cautionary approach.

If I could just add in a little bit more about your question about the AI policy.

And I just want to say that at our campus we did talk a lot about AI policy and there wasn't a general agreement about what exactly that AI should look like that would be a one size fits all for the entire University.

But what we do offer our faculty is something we developed as an AI policy builder that each faculty can use if they wish to create clear usage guidelines about what constitutes reasonable use in a course for students for use of AI.

So that because it's very different how you might use AI in an English course course versus a math course versus another course. So I just wanted to put mention that as well, that I think that there's still some thoughts developing around AI policies and what makes sense as a campus based approach.

And so in the meantime, those are some of the things that we're doing now.

I appreciate that. I'll say that I aged out of my children's math many years ago and that has been challenging in my household. I'm a more liberal arts person. And recently my daughter was asking ChatGPT not to do her problem but to teach her. And it was like, step by step, how do you do this?

And gave it different numbers so that she could then do her problem. And it was super helpful to our family. Now, I did then use a calculator, make sure that it was right, but I do think in that context, she was actually using it as a learning tool. Right? Not a cheating tool. And it's an important distinction.

And our young people are much more sophisticated at their use than we are, and I think can use it in ways that helps them without supplanting learning, which can be really great. So my last question is, how will we know if this is a success? Are there metrics? Are we watching? Are we learning from it?

Are we going to, I hear a lot of we're going to iterate on this, but what are the metrics to decide if what we've done is the right move for the CSU students?

You are referring directly to the metrics effort that is underway right now. We're collecting metrics from, I believe it's nine different lenses, ranging from adoption to workforce employment outcomes to even, you know, we're actually asking on academic outcomes as well, environmental impact. Those are all lenses that are part of our effort in this metric space.

We want to measure how we're doing. We want to know the good and the bad, and we want to track that over time so we can improve. And as you pointed out, you know, we understand all these criticisms and concerns, and they're valid. They're valid concerns.

And the best way to deal with those concerns is to have our universities participate in helping to shape the future of these technologies. We can't just sit back and let it go by. We should help shape the future.

And I know that we're going to hear this in the next panel, but I did have one more question before they come, because you will all have left us. I am a big believer in the path forward in California, and I said this in my opening remarks, is students and teachers and workers at the table.

I think that that is so critical to doing this right and doing it in a way that moves us forward rather than leaves us behind. You mentioned they were part of these working groups. I appreciate that. I want to make sure we're not tokenizing that.

Can you talk about how you're making sure that their input is really meaningful?

Well, they. So the working groups, the main system wide working group has membership from all those areas that you mentioned, faculty, staff and students. And they are just as much. Their input is just as much valued and utilized as anybody else in that group. In fact, their recommendations are the report that you saw or you've read.

And then we continued that, that Committee has reconvened and we've actually added more students and faculty and as as many as possible, as many who are interested, who have stepped up and asked to be part of it as well. So we're hoping to hear even from new, new voices or a larger variety of voices.

Great. Thank you so much. Turn it back over to the Chair.

Thank you so much, Madam Chair. And thank you to everyone for your comments on this. Just to echo the number of the comments around uniform training system wide training, I think those are opportunities going forward that we should definitely look at and look forward to report backs on that.

One final question I had on this and this is touched upon this earlier, but the recent MOU between the CSU and the Governor calls on AI Workforce Acceleration Board to advance workforce development initiatives that align the CSU's emissions of equitable access, economic mobility and AI related opportunities.

You touched upon a number of the things that you said, internships and other areas, but are there any additional initiatives that are in development on this space?

Some of the things, the Workforce Board is, you know, just voluntary members that have come together to help, you know, shape how we might do student prep in this rapidly evolving workforce setting.

So some of the things that the members have talked about is, for example, creating some sort of index, workforce readiness index where we, students can demonstrate they have the durable skills like creativity and collaboration and communication in connection with these technical skills are going to be going to be important with using these tools and also being able to leverage them in their context.

These AI tools are going to be not only in STEM fields, but they'll be using creative fields as well and those kinds of things. We're trying to capture all of that. The micro internships we're doing a pilot this fall and the AWS innovation event that Leslie and I attended.

It was really just a stunning thing because we had students from every campus in the CSU, many of whom who had never used technologies, come in, learn these tools and actually work on a real world problem. And it was. And they were delivering real working models. Those are just some of the early wins as this is going.

We're just about to have our third meeting ever of that group and we.

Continue to explore future opportunities with all of those vendors and we're excited about that.

Thank you so much. Did you want to add to that?

Oh, I was just going to say in San Jose State, we're doing a lot around in that space. I mentioned the Hackathon event that we're doing with IBM, which is providing opportunities for students to get mentored by IBM professionals. And we have other kinds of micro internship opportunities that we're exploring. Adobe Ambassadors.

There's a lot of things that we're doing to help prepare our students for that.

Great. Thank you so much to Dr. Clark, Dr. Kennedy and Dr. Hirsh for your your work and efforts. And thank you to Ruthie as well. So appreciate the insights shared by the first panel. Next up, we'd like to introduce our second panel.

Our speakers for the second panel include Dr. Elizabeth A. Boyd, the Chair of the Academic Senate of the California State University, Dr. Elaine Bernal, a Member of the Board of Directors of the California Faculty Association, Catherine Hutchinson, President of the CSU Employees Union, and Katie Karroum, Vice President of System Wide affairs for the California State Student Association.

Each speaker will have six to eight minutes to provide your respective testimony. Thank you so much to each and every one of you for being here today, and you may begin when you're ready. Welcome.

All right.

Good afternoon.

Good afternoon. Do I need to enable the mic?

Maybe move it a little bit closer.

Just didn't want to breathe too hard into it. Good afternoon, Chair Fong, Chair Bauer-Kahan. The members of the committee, thank you for being here. My name is Elizabeth. Some of you know me as Betsy or Liz or Dr. B. Boyd.

My pronouns are she, her, they, them. I serve as the Chair of the Academic Senate of the California State University, otherwise known as the ASCSU, which is the system level non bargaining co-governing body representing nearly 29,000 faculty of the now 22 campus California State University system.

My home campus is Chico State, where I've served for over 17 years as Professor of Entomology in the Plant and Soil Science program within the College of Agriculture. And I'm proud to sit here before you today as the first Professor of Agriculture to serve as the Chair of the ASCSU.

Let me begin first by thanking you for creating space and opportunity to speak with you today. And thank you also for making space for our labor colleagues and for students on this important topic.

Each of them will raise important points for your consideration and I will do my best not to duplicate those and do want to express my support for the points that they will share with you today.

So, I'd like to provide some background and clarity on the faculty involvement in the realm of AI in the CSU system from a faculty perspective. And I thought it might be best to address this in sort of a timeline format.

The faculty at CSU have been watching and engaging with various AI entities and the corporations associated with them for nearly a decade. And the ASCSU has issued at least seven formal recommendations and communications on this subject since 2019. A subset of the most relevant I will cover today.

And I've also provided those to you as background information which just went forward today. So, I know that those will come out for you soon.

As you heard in the first panel, the ASCSU has been involved in sort of sounding an AI alarm, if you will, for several years prior to the most recent MOU and contracts, culminating in 2019 with our first formal communication on AI calling for a working group on the impact of AI in higher education.

The ASCSU recognized the need to address the revolutionary changes and potential impact of generative AI. This call was renewed in March of 2023 and the CSU responded by establishing a Generative AI Committee on which then the then chair of the ASCSU and another senator served on this working group.

In 2024, the ASCSU resolved to recognize the current system wide AI professional development opportunities and highlighted the AI Commons. They also called for additional AI teaching and learning and faculty development.

Furthermore, in 2024 they passed a resolution requesting more representation on the AI committee and finally a resolution that was issued concerning student intellectual property and the potential use of GenAI detection tools and student course assignments.

In that, the ASCSU urged the Office of the Chancellor at the time to fund professional development courses in AI to enhance ethical AI tool usage and that not only in the college classroom but by faculty and students universally through this CSU.

They suggested that the teaching and learning courses would include but not be limited to, you know, what is AI, where is, what is it, what it is not, what it means for the college classroom, the AI detection software, how it can be used, what it can and cannot do in terms of student success, creating campus community based aims and values for AI use, enabling ethical student use of AI generated content and enabling faculty and or faculty and student co development of AI classroom policies according to course content by discipline.

And they asked for additional resources and for adequate funding of these ventures to ensure that faculty had an opportunity to participate. So, despite sort of the instigation and involvement in the AI space, the January 2025 news of the contract with OpenAI and the development of the CSU Workforce Acceleration Board did come as a surprise.

There are a number of lessons that we could learn from the communications surrounding the rollout and the apparent endorsement of AI to our students through the agreement and issuance of license without ensuring all faculty were trained.

So, I think that these concerns have been brought up with some of the questions about the pilot, and in step with that announcement though, the ASCSU was invited to place senators on a refreshed Generative AI Committee, of which we have appointed several faculties to each.

I also serve alongside another senator on the Workforce Acceleration Board and each of these committees have several subgroups or working groups that range from teaching and learning to procurement and purchasing.

In my role on the Workforce Acceleration Board, it was incredibly refreshing to hear that our industry partners on the board and share some of the same concerns that faculty have about issues of access, scale and developing critical thinking, and the reinforcement of ethical use.

They also see incredible opportunities for our graduates and are helping create the AI Literacy Index for our graduates as well. As this work is still in its infancy, the ASCU has also been working on additional formal recommendations and issued two of those in May, one of them on the possible use of AI in instruction and then also urging caution in the use of generative AI.

And I'll just read a couple excerpts from those, which is that the ASCSU was urging that the CSU consider other challenges surrounding GenAI, including that as a technology, it's still in its infancy; it's currently sometimes unreliable or regularly produces inaccuracies or has bias, sometimes often hilarious and other times a little bit disturbing or wholly disturbing.

The commitment of about 17 million in financial resources by the CSU did come about without consultation and there's concerns about its sustainability, especially in our pedagogy. The text, digital imagery, audio, video, and other information and products produced by GenAI are often built on the intellectual property of others, frequently without remuneration, citation or even recognitions.

This is an issue we hope that we could dive into. The enormous use of environmental resources was also something we wanted to be cognizant of, particularly in the energy needed to train these GenAI models.

And lastly, possible exploitation of low wage labor training and the running in GenAI models, and replacement of humans in much of what we do in a front facing CSU to our students.

Overall, there's a smattering of faculty involvement across the CSU and the AI Commons is a huge step forward for our system, offering self-guided courses and certificates.

While still more of our campuses have been working on AI at variable speeds and adoption from the lens of the ASCSU, there's both widespread excitement and extreme caution that I know my colleagues here will address as well.

To be successful, we need funding and support for faculty trainings, like those offered at the AI Commons, and at faculty development centers at each of the campuses. All faculty have access to the AI Commons, but as a system these may not be as comprehensive for all the disciplines that will be trying to work with these tools.

And having dedicated time means that we need money to do that as well. So, Chairs Fong and Bauer-Kahn and members who stated concerns that are also faculty concerns, and there still are new concerns as we learn and move forward with these tools and how we integrate them in the classroom.

We want our classrooms to continue to be safe learning experiences for our students. The biggest concern for most faculty is how to proceed.

Ensuring that we continue to support the diversity that is the hallmark of the State of California and that these tools are used with a conscious eye on equity inclusivity and that we maintain high fidelity in the ethical use of these tools while ensuring that we maintain a learning environment where students have the ability to critically evaluate the world.

Ensure that the CSU continues to maintain a degree that elevates the social and economic mobility of our students, that our graduates can contribute and elevate an inclusive democracy and succeed in their careers, that academic freedom and intellectual property remain in the hands of faculty and the students that employ the tools.

And lastly, that AI is not a substitute for human or human experience. Thank you again for your time and for calling this important hearing. Your engagement is essential to the success of the CSU, and we really deeply appreciate that.

Thank you so much. Welcome.

Hi. Good afternoon, Chair Fong, Chair Bauer-Kahan, and members of the committees. Thank you for the opportunity to speak today. My name is Dr. Elaine Villanueva Bernal. My pronouns are they, them, theirs.

I am a lecturer in the Department of Chemistry and Biochemistry at CSU Long Beach, and I serve on the Board of Directors of the California Faculty Association as Associate Vice President of Lecturer South. I am also proud to be the first Filipino person to serve as a board officer in the CFA's 43-year history.

Filipinos have a deep and powerful history in California's labor movement, and I carry forward that legacy of collective struggle for dignity, equity and justice. CFA represents 29,000 faculty across the CSU: professors, lecturers, counselors, librarians, coaches serving half a million students every term.

I speak to you today with both the lived experience of a faculty member in the classroom for nearly 20 years and with the responsibility of the collective voice of our union. In chemistry courses for non-STEM majors, I see students are anxious, convinced they did not belong in science.

Chemistry is one of the most historically difficult courses, with high failure and drop rates, and nearly every student has had a cautionary tale about it. But in my own classroom, we have already been using AI thoughtfully with ethics, equity and accessibility at the center.

When framed this way, students begin to connect chemistry to their own fields, from art to business to health, communication, and even dance. They discover confidence and ownership in a subject they once thought was closed off to them.

If students can find that agency in chemistry, one of the most challenging courses in higher education, then AI has the potential to help every CSU student feel capable and included. But AI also risk undermining critical thinking.

If students come to believe that they can outsource analysis and reasoning to an algorithm, they lose their very habits of questioning and reflection they that are core of a university education. One lesson I carry from my work is that true student success depends on autonomy and agency, on students having the ability to shape their own learning experiences.

Protecting privacy is the foundation that allows students to trust the system, take ownership of the education, and engage fully as partners in the learning process. The same is true for our faculty. Our success and academic freedom depend on the professional agency in teaching and mentoring, supported by tools and policies we can trust.

I know this not only from my classroom, but also from my research. My doctoral work on the CSU's course redesign with Technology program found that the most meaningful improvements did not come from simply adding tools. They came when faculty collaborated, redesigned intentionally, and used technology to deepen rather than replace human beings in teaching.

Those redesigns led to measurable GPA gains and student engagement. When both student and faculty agency were supported, innovation worked. When either was ignored, it failed. This is why this AI initiative raises alarms. It was rolled out without consultation with faculty, campuses, or our union.

Workplace conditions are part of our collective bargaining agreement and any change in workload is mandatory subject bargaining under HIRA. The Chancellor's office has a legal obligation to meet and confer with us. Instead, this initiative was imposed unilaterally. The dollars are staggering. The CSU has already spent $17 million in this initiative.

At the same time, Sonoma state faced a $24 million deficit that led to the largest layoff in its history, with faculty cut, departments closed, and athletics eliminated. This juxtaposition reveals CSU's budget priorities, flashy tech and corporate partnerships over the livelihoods of workers, and the stability of our campuses. Faculty and students deserve better, and legislators deserve budget transparency.

And yes, the CSU has invested $3 million in competitive grants with awards $30 to $60,000 each to spark faculty innovation. That is an important step, and many colleagues have developed creative projects through this opportunity. But grants like these are the beginning, not the end of what is needed.

California's workforce will not be shaped by handful of pilot classrooms. It will be shaped by every CSU graduate. The challenge now is scaling those promising models across all 23 campuses and disciplines. And this is where the Chancellor's Office falls short; presenting competitive grants and vendor partnerships as if they constitute a system wide strategy. They do not.

We are also deeply concerned about job security. Managers have already tried to roll out AI tutoring bots that hallucinate and misadvise. Our union had to intervene and to shut them down. Bots cannot replace faculty, counselors, coaches, or librarians. There are risks around intellectual property.

It would be naive to believe corporations provide AI access for pennies without extracting value in return. What they want is our data, student queries, assignments, and faculty teaching materials.

As Ethan Mollick points in CO Intelligence, AI companies are running out of high-quality data to train on, which makes our classrooms especially attractive sources for knowledge and data extraction. These systems can also be weaponized for surveillance over faculty, assign and teach, especially those of us teaching race, gender or power.

That is a direct threat to academic freedom. And let's be clear about bias. AI does not just make random errors. It reproduces structural racism. As scholar Ruha Benjamin reminds us in Race Africa, these systems are not neutral. They encode inequality, including anti blackness into their design.

CSU students, Black, Indigenous, Latinx, immigrant, queer, trans, first gen, disabled students should not be handed tools that reinforce the discrimination they already face in society. Which is why it is ever so crucial to ensure the longevity of our ethnic studies programs and departments. The environmental tool is also real: generative AI consumes staggering amounts of electricity and water.

Every quarry draws on server farms cooled with water diverted from communities already facing scarcity. Data centers now account for more than 4% of the U.S. electricity usage. Much of it's still fossil fuel and training a single large model like GPT3 can consume up to 700,000 liters of fresh water. CSIR proclaims the sustainability mission.

Yet this initiative accelerates environmental degradation. Training the next generation is meaningless and if they do not have a sustainable future to inherit. AI is also creating more work for faculty. We are redesigning assignments, rethinking exams, adjusting classroom policies, and double-checking work with unreliable detectors, all on top of existing responsibilities with no added support.

This is why CFA is calling on CSU managers to consult faculty and staff first. When AI changes how we teach, how we grade, how we advise, it changes our working conditions. CSU must meet and confer with our union as the law requires, provided funded, optional training at scale.

If faculty are to integrate AI thoughtfully, they need time and resources, stipends, release time, professional development, not unpaid labor. Keep humans at the center. Students come to CSU for mentorship, guidance and care. AI can supplement but must never replace this. Protect privacy and academic freedom. No faculty material should be uploaded into corporate systems without consent.

No student data should be mined or sold. Protection should meet the standards of the General Protection Regulation in Europe and California's own Consumer Privacy Act. Require equity and bias audit. Independent audits must be mandatory before CSU deploys any AI tool. Communities most at risk are the very students CSU is meant to serve: Black students, Indigenous students, Latinx and immigrant students, queer and trans students, and students with disabilities.

Mandate transparency: if CSU signs contracts with tech companies, those contracts must be public. Tie AI budgets to human support: every dollar spent on AI must be mandated by investments in human capacity, faculty hiring, advising, counseling, and student support.

And finally, ensure budget transparency. Legislators, faculty and students must see exactly how AI dollars are being spent. Faculty and students must not be treated as beta testers for corporate tools. If AI is truly empowered, it must reduce inequity, respect privacy and put humans at the center.

If my non-STEM students can find confidence in ownership in chemistry, again, of course with one of the highest failures in drop rates, then every CSU should be able to find the same sense of agency in their education. The question before us is simple: will AI help us build that future or will it erode it?

True innovation in public higher education is not measured by contracts or pilot projects. It is measured in whether every student and every faculty member is empowered to succeed. That is the standard the CSU should meet and that is the standard that California deserves. Thank you.

Thank you so much. Welcome.

Thank you. Good afternoon. Chair, Co-Chair and member. My name is Catherine Hutchinson. I'm President of the California State University Employees Union and I speak to you today from the perspective of the staff of the California State University.

We are the people who keep this CSU running every day through student services, IT support labs, libraries, payroll facilities and beyond. The CSU recently made a significant investment in artificial intelligence. For many, AI feels like the beginning of a revolution. But here's the truth. With any revolution, technology only succeeds when workers shape it.

Without the people who understand how campuses function, AI risk being another top-down conceptual rollout that disrupts through fractures rather than improving higher education. When systems like AI are introduced without input from staff, they fail to account for real working conditions. Workers need a seat at the table not as afterthought, but as a central partner.

We are the ones who understand the outlier scenarios, bottlenecks the student and faculty needs and which gaps technology might fill or create voids. We are the driving force that will make GenAI a reality. If AI is indeed a revolution, a successful implementation goes beyond a contract, handshake software licenses, a couple of consultants and non-CSU pre trained modules.

It involves organizational restructuring. That restructuring must be intentional, negotiated and just. We have some solutions which are create worker councils where employees and management collaborate in a visionary manner on how in which AI tools are introduced, tested, and refined for successful implementation.

Negotiate guardrails to ensure AI is used responsibly, protecting data privacy, our intellectual property, preventing bias, and avoiding job displacement. Reducing worker anxiety by involving staff early and often and empowering them. Our employees should not encounter surprise, steep learning curves or feel replaced by machines, but instead be empowered by them.

And of course, focused adoption that improves productivity, and student success rather than chasing flashy tech trends. And critically close surveillance of this technology. AI is only as strong as its weakest output. We all know these systems can generate nonsense or incorrect results if left unchecked.

If we're going to rely on AI for advising, records, or even payroll, then oversight by skilled workers is essential. Staff must be given time and authority to monitor the technology, identify and correct errors, and improve its accuracy before any discussion of replacing human worker occurs.

Job displacements should be held off until this technology is not only proven, but also reliable, tested in real conditions by the very workers who understand these conditions best. We cannot ignore that AI adoption will cause disruption. Some jobs will change. Some tasks may be automated. For this reason, we call for a just transition.

Proactive strategies that help workers adapt and thrive. That includes severance protections. If jobs are to be eliminated, justifications must be laid out, so families are not blindsided. Profit reinvestment incentives that retrain and redeploy staff into new roles created by AI, and shorter, more productive work weeks as workers navigate retraining and disruption, ensuring no one is left behind.

Like the CSU, the staff are also generational, with many of them being graduates of the very campuses they serve. We know this AI revolution will be generational. This is why we stress our technology collaboration. Our CSU staff are in the forefront of innovation. We're the ones who make it work.

But the lesson of every technology revolution is this: progress is only progress when workers and communities are lifted up with it. If the CSU wants AI to succeed, then workers must have a permanent seat at the table.

Together, through worker councils, negotiated protections, close surveillance of the technology, and just transition strategies, we can build an AI future that strengthens our university, respects our employees, and truly serves California students. Thank you for your time and your commitment to ensuring that CSU's AI's future is a fair one.

Thank you so much. Welcome.

Thank you. Good afternoon, Chair Fong and Chair Bauer-Kahan and members. My name is Katie Karroum, and I'm a student at California State University, Northridge. I also have the humble opportunity to serve as the Vice President of System Wide Affairs for the California State Student Association.

CSSA serves as the official voice of the over 460,000 students in the CSU system. And I appreciate you taking this time to hold this hearing and allowing me to share how the CSU's artificial intelligence empowered initiative is impacting our students. I want to begin by recognizing the opportunities that this initiative creates. The launch of ChatGPT EDU represents an innovative milestone for the CSU and will be highly beneficial for students.

By implementing ChatGPT.edu system wide, the CSU is ensuring that all students will, regardless of campus resources, benefit from cutting edge AI technology that enhances both academic and professional success.

Without this system wide initiative, access would be limited for students at campuses unable to afford individual licenses, creating disparities in opportunity that this initiative is aiming to eliminate.

This system wide approach ensures that every student, regardless of which CSU campus they attend, has equal access to the AI tools and experiences necessary to thrive in a rapidly changing digital world.

Through a call for faculty proposals and a $3 million in grants to support ethical and creative AI in the classroom, the CSU has shown its commitment to preparing students for a rapidly evolving digital landscape, which is imperative.

Students cannot afford to be left behind as AI reshapes the workforce, so this initiative helps to ensure that CSU students are ready to succeed in their chosen fields post-graduation. With that being said, students are experiencing ever growing challenges and voicing their concerns when it comes to the rollout of this initiative.

It is great to hear that 93,000 accounts for ChatGPT.edu have been activated. However, there are valid concerns about how many of these account holders will actually complete the AI training needed to use the tool effectively.

Ensuring that students, faculty, and staff not only have access, but also the knowledge to integrate AI into their learning and workflows and will be critical to realizing the full benefits of this system wide initiative. Most importantly, staff and faculty being properly trained on this could directly influence student success inside and outside of the classroom.

While the Chancellor's Office has begun providing system wide resources, students are increasingly concerned about how these supports will translate into meaningful implementation at the campus level. We appreciate the ChatGPT Campus Ambassador Program. However, it is important that the promotion of OpenAI resources does not fall solely on these students.

To ensure full accessibility and transparency, these resources should be embedded within the programs and support services of each CSU campus. Some professors openly encourage AI literacy and integration within their classrooms, while others penalize students under suspicion of AI use. They do this through the utilization of AI detection tools that have been deemed unreliable.

In some cases, students have even faced mass accusations. For example, one professor reportedly turned in 50 students for alleged AI usage. Many students worry that their academic credibility may be tainted if professors assume that any polished work must come from AI.

This also affects their time to graduate, as several students have received zeros on assignments due to AI detection. Whether the detection of AI is false or accurate. Some faculty have shifted away from written assignments to oral presentations or other formats.

To avoid any AI concerns, we encourage faculty to embrace multiple teaching modalities that reflect the diverse ways students learn best. However, we also believe this approach should extend to the way students' knowledge is assessed.

Classroom environments must be curated as safe spaces where students can focus on retaining knowledge and demonstrating their learning from rather than feeling as though they may be penalized at any given moment. Overall, this kind of inconsistency leaves students uncertain. Are we allowed to use AI or not? What counts as ethical use?

These types of inconsistencies within the classroom when it comes to the usage of AI are not sustainable, and students have growing concerns on how this initiative will alleviate those inconsistencies or exacerbate them. Students also have concerns about data privacy.

While we were told that ChatGPT keeps our data private and will not be made available to others, students still have questions. We want to know: what protections exist for our conversations, assignments, and personal data. What happens after the contract expires? Will our data be retained, transferred, or subject in renegotiation?

Who is responsible for monitoring compliance with data protections? The Chancellor's Office or individual campuses? Students want clear, accessible information on AI data policies in plain language, not just in legal terms buried in contracts. Many students at the CSU are very conscious of the environmental impacts that AI has.

Training and running AI requires substantial amounts of energy and water, raising questions about how much of these resources are being consumed to support AI across the CSU system and what the implications are for the CSU's broader sustainability goals.

As climate change intensifies and California faces increasing pressures on its waters and energy systems, it is important to weigh the environmental tradeoffs of adopting AI technology system wide. AI adoption should not be considered in isolation from the CSU's existing sustainability commitments, including its Climate Action Plan and carbon neutrality targets.

If the CSU is investing in technologies that significantly increase energy and water consumption, this could directly conflict with its goals of reducing greenhouse gas emissions and building a more sustainable future, which are all areas of concerns that our students feel strongly about.

Students urge the CSU to report on the carbon, energy, and water footprint of the AI tools being used system wide. The CSU's AI initiative is ambitious and innovative, and students want to be a part of shaping it.

I'm proud to serve on the CSU Generative Artificial Intelligence Committee this year, and I hope to continue sharing the student perspective in that space as we shape system wide AI policies.

Legislatures: as you continue to examine this issue, I urge you to hold CSU leadership accountable to ensure consistent, transparent campus policies, clear safeguards for student privacy and data, assessments of environmental and ethical impacts and, above all, meaningful student input at every stage of implementation.

Without clear guidance from the Chancellor's Office and with campuses left to act independently, students will shoulder the burden of navigating how to implement this initiative while system wide perspectives on AI remain fragmented. It's imperative to properly assess this initiative in ways that do not put students at the crossroads of the implementation process.

The CSU faces a defining choice. Will we be AI powered or truly AI empowered? The difference is seismic. If the CSU and the legislature are serious about harnessing artificial intelligence, they must deliberately invest in funding that drives student opportunities, not just the contract with OpenAI. This initiative could reshape higher education as we know it, but that transformation will only happen if students are at the center every step of the way.

I want to thank you for holding this hearing and for centering student voices in this conversation, and I'm happy to answer any questions that you may have.

Thank you so much and thank you to each and every one of you for your testimony. Powerful insights here today. Actually, we'll start with our CSA rep. Thank you for your testimony.

You mentioned in your comments that there needs to be additional training, there needs additional support, and also using AI in an ethical manner to support your education. Has CSU, in the months that you've been on this Workforce Acceleration Board, have there been additional consultations and do you see additional progress in this space?

In the training aspect?

Training aspect, and your involvement also with the Workforce Acceleration Board?

Absolutely, yeah. I just recently got appointed to that committee within the last month or so. But really the focus from my perspective has been on policy and setting metrics, as opposed to pushing training or making sure training is going well, again, from the perspective of what I've seen in the committee so far.

So I think it's been more foundational work.

So, in terms of that foundational work and the policies that are being worked on, what more can be done to support CSSA? I know you mentioned a number of them in your remarks.

I mean, I think it really is just more student involvement in policymaking and more system-ness. I think something that I think the Chancellor's Office does well and can be interpreted well from campuses is allowing autonomy for campuses to operate how they will because they know their students better than the Chancellor's office will.

And I think with something like this, there's this idea of there being, you know, let the campuses kind of, you know, figure out how they want to roll out the AI with their students. But I think that will inherently create some inequities. If we're having campuses that are being more forward with their faculty about training students and faculty on AI ethical use and just proper usage versus other campuses not.

So, I think encouraging that to the Chancellor's Office and to campuses, in general, is going to be very imperative. Yeah.

Great. Thank you so much for your comments there. And to the Academic Senate: you mentioned also the representation on the Workforce Acceleration Board; in the comments here. as the board is meeting more regularly, are you more fully included in those conversations and shaping those policies as well and what more can be done there?

Yeah. So, I have, as a chair of the ASCSU, I feel I've been very, very included since the development of it, since it's been - since the board has been meeting, been able to be at each meeting. There are subgroups that are currently working.

In fact, just this morning at 10, I was working on a subgroup, the Workforce Acceleration Board. And we're working on how do we scale the work that we need to do.

Much of the examples that Vice President Karim mentioned, the one offs with the faculty and how they're utilizing AI in the classroom or not, and how it's being interpreted, it's critical that we have an educated workforce in the CSU currently, while we're also trying to educate our workforce in the State of California.

But to your point, I do feel that there's been involvement. In terms of what we could do better. I would say it might be useful. And I do think there's an openness to expand that to faculty with specific skill sets, as they have been doing with the Generative AI Committee that has been very expansive and open.

I know that the workforce one has been a little bit more dialed in, but I do think it could benefit from a few more voices there.

Thank you for that context. Sounds like some additional voices on there, additional representation to make sure that all voices are being included in this process would be critical. This question is for CFA. Thank you for the context here as well.

Thank you.

We know that CFA filed an unfair practice charge alleging that the CSU violated the law during the rollout of the initiative by failing to meet and confer since then. Has the CSU increased the consultation with CFA since that rollout?

No. And I think it's an opportunity to be more involved. As I was listening to CSSA and Betsy speak, and thank you for your remarks.

I see this as an opportunity for CFA to be a convener of faculty, to hear all voices, to hear all areas of expertise, because I think it's not just having the AI expertise, but what does that look like in specific disciplines? What does that look like interdisciplinary? Right. I saw it in my own class.

I had a student who was a dance major and, came up with a final project of here's how to use dance to explain how molecules move. I mean, how amazing is that? Right? But that's just my classroom.

How do we come together, especially with CFA being a convener with all our 29,000 members and hear that feedback and see how we can best serve the students of California?

Thank you so much. Sounds like the increased consultation. There's going to be a lot more communication and consultation going forward. So, I appreciate the context and the insights there. Thank you. I'll pass it over to my - I'm sorry. Assemblymember Patel.

Yes. Thank you for your presentations and your thoughtful insights. It's always good to hear the full story from multiple perspectives. That helps us understand the challenges a little bit better. I'm going to get back to a question that I asked the administrative level panel before.

When we're looking at policies, specifically around AI use policy, what I've heard from several students in my district is that, and not just from San Marco, CSUSM, but just SDSU and students that attend other colleges and universities around, that faculty don't always have an AI use policy, but yet will have a discipline policy.

And so, what students ask for is, can every faculty have an AI use policy, whatever that may be, and just make it very extremely transparent, so they know what their expectations are. Is there - first of all, does that sound consistent with what you're hearing from your peers?

Totally. Yeah. I would say I don't even know if - again, from my experience as a student and from my experience speaking with my fellow students, discipline definitely ranges depending on the professor, and there isn't clear policy that comes out of faculty affairs. And I used to be like Student Body President at Cal State Northridge last year.

And I worked very extensively with our faculty senate. And this was still a thing that was coming up, where we had professors on one end that were, again, very encouraging of AI usage and we're integrating into their coursework. And then on the other end we have professors that are, I don't know where they get this idea of saying, like, you know, "I'm going to use these three AI detections tools, and if all of them give me above a 60% AI detection that, I'm going to fail you on this assignment."

I don't really know where it comes from, but AI use policy, I haven't seen it, but I think that would be much more efficient and transparent for our students.

And certainly, the challenges with the AI detection tools are it flags good grammar, oftentimes, and some students just have really good grammar. We hope that they would by the time they reach university level.

Agreed.

And sometimes those algorithms are around good grammar.

Yeah. And our students are smart, so I don't blame them. Yeah.

And sometimes also, I've seen heard stories from students in the system where they've produced a narrative that is very creative and something unique, and it's been flagged as being something too creative and unique, and it must have come from AI. And so, there are definitely challenges around that to prove innocence.

But also, just in general, to the faculty on the panel today: is there a possibility to get to a point where all faculty are providing some kind of AI use policy just so students know what the ground rules are in each classroom that they walk into?

It would be lovely if there was one general one, but I think we're not going to get there. Faculty have the right to determine their own grading environment and standards and expectations, and we would want them to have that autonomy. But having some kind of expectation seems like a fair ask.

I think so, too. I think fundamentally, many of us in education, I think education, in general, needs to move away from a punitive model of teaching students. Like, I've been teaching for almost 20 years. I'm still here to learn. I'm learning a lot right now. So, and this is where academic freedom comes in.

You know your classroom best. You know your students best. It is your relate - it's between you and your students. You craft that working relationship. I had a colleague who is in Philadelphia, so what he did was that him and his students actually came up with the policy together in their classroom. So, it's something where you can, yes, generally, "Hey, you want to do something with your students?

Here are some guidelines." Right. So, every semester, because every semester is very different. I have 210 students a semester. It's a whole different audience every 16 weeks. So, it's a way to have community and agency within the classroom.

Yeah, for sure. Thank you for the question so much. So ASCSU actually recommended syllabus language for the system on AI, and I'm not sure how that's trickled into everything yet, but it's out there. And certainly, you know, there's no harm in recirculating that.

Also, faculty have been involved at various levels at the campus and system level, in terms of some of the policy development that you're talking about. Some of it is housed in digital learning policies.

So, it's not just AI but wholly encompassing of all the digital tools that we use, and particularly those that are the detection tools, which are very much a concern for us right now.

And lastly, I just wanted to mention that some of what we're seeing here is that with the adoption of this technology and as we move forward, there's always different levels of acceptance and use as we go. But I think another sort of overarching concern is the number of students in the classroom.

The CSU prides itself on its student to faculty ratios and its student to faculty interactions. And when we lose that ability to have that interaction because our classes are too big, then we fall to these other tools that are, that have problems. And often it's the same thing for students.

When students are overwhelmed, which many of our students are coming into the CSU, they have jobs, they're supporting families. There may be not a traditional student in the sense that we think of them. They need resources, and it's easy to fall into a quick use of a tool rather than going through that process sometimes of developing that.

So, I think that we're seeing sort of a culmination of a lot of that that I don't - I think our policies could really help us with in terms of guidance, but there's some other gaps that I'm not sure that can be solved in those ways.

Not that you were trying to suggest that, but I think they overlap and intersect.

Yeah. And certainly, when I was in college taking a higher-level calculus class, we had a retired professor come out of retirement to teach our class, and he said, "You can't use calculators because the human brain invented calculators, and you have tables and charts that you could use to do all your calculus."

I don't want us to have that kind of mindset as AI. We have this tool in front of us. It is a powerful tool, but we should definitely produce and maintain and even tweak those guardrails as we move forward with how we're using them. But I think student voice has to kind of lead the way. Yep, absolutely.

I really believe that. I think our students really, they know the technology that they're interacting with. They are part of its development and growth. They're the largest user of AI tools, and I think we need to let them direct how they want their education delivered to them.

Thank you.

Thank you so much. Chair Bauer-Kahan.

Thank you. And I have to say, Vice President Krum, I don't know. We're doing great at the CSUs. This is. Your testimony was so impressive. Maybe we don't need to change anything, and that's where broke things. But I think I really appreciate the conversation that Aslia Patel led around policies.

I was alarmed, frankly, that there wasn't much direction coming down, especially given the way the rollout happened, where they were hitting handing this to you, and then it didn't sound like handing you a lot of helpful guidance, if you will.

And I think I would say I appreciate what you're saying about academic freedom and policies being set at the classroom level. I also think model policies like you're mentioning in the syllabus could be super helpful. This is complex stuff. I think I imagine a world where they're right.

It would apply differently in an English class than a math class. But you could have math faculty coming together across the system to come up with a model policy that, again, can be tweaked at the classroom level but provides some starting point, because I don't think every faculty across the system is prepared to do that.

And that's also okay. So, you know, I mean, I think people have different literacy levels at this point, and we should be supporting them in this adoption, but again, also balancing with freedom. I also was alarmed, as you probably could tell, by what seemed to be inconsistency in the training.

I think, you know, and I heard you say this, I think both of you say this, that you could sign up and there was no obligation to actually take the training, which was wild to me. I'll be Frank. I think that if we're handing you this tool, there should be some expectation that goes along with it.

I also will say, I think I would like to see the system say there has to be a policy, which I think is what Senator Purcell is saying. Again, there should be model policies. I think the faculty should be able to tweak them.

But I think that if a student is being reported, as you mentioned, there should be no discipline unless it was clear to the student what the expectations were. And that period, like, I don't know how we can discipline students when they don't know what the expectation is. And it is different classroom to classroom.

And it can be different classroom, classroom. If they walk in and they are told clearly what those expectations are. Our students are smart. They can meet different expectations in different classrooms if they're laid out. And then I think the discipline should only be dependent on what they were told. That.

That I just wanted to say because I was upset to hear that. I think that students should. And I will say that when my son was flagged by one of these AI detection tools, I said, that's awesome. It means your grammar's really good. Because I had watched him write the paper, so I knew he didn't use AI.

I mean, these tools. That's exactly what they're doing. I was like, wow, you must have incredible grammar. So it isn't fair. And one of the things this Legislature has been pushing through budget actions in my six years here is moving to 100% of our CSU students graduating in four years. Those numbers need to change.

It is costing our students too much in time and money to not be able to get through our CSU system as quickly as they would like. And so if this is an added problem to getting our CSU students to graduation and onto the lives they deserve to lead post degree, that needs to change, period.

Because I know that is a priority of this Legislature. So I wanted to add those comments on. I am going to ask one of. And I don't know if this is the time to do that. I was going to ask Mr. Clark to answer one of Vice President Croom's question.

I don't know if you'll come up, Mr. Clark. I thought it was a really good question that was posed. You were very clear in the first panel that during the course of the contract, the information is private. The question was posed by the Student Association. What happens after the end of the contract?

Can you answer that for us? Yeah, I thought it was a great question.

Yes, it is a great question. I don't know if it's explicitly called out in our contract, but our understanding is that data would just expire. If the contract expires, it's just going to go away. There's no way to take the data with you when you go, when you leave. Yeah, it doesn't.

Yeah.

So I would ask you to follow up with both of our committees on whether that is explicit in the contract. It should be. And if not, I would ask that you change that. I don't know if you want to.

Yeah, please. Definitely would love to see some follow up on that, please. Okay.

Thank you. And then. I really appreciated, you know, Chair Fong raised the question of bias in the first panel. I think it's a really important one, as they say, as it relates to AI. Garbage in, garbage out. And so these tools, although powerful and impressive and Wonderful.

Can also output a lot of garbage, which I think you, Professor Bernal, were alluding to in your comments around protecting our students and especially our students of color and disabled students and our students who are from protected classes. And I really appreciate that.

I watch my kids usage of these tools and like many young people, he's navigating different tools, trying, you know, that's what they're doing and it's great and they're learning from it. And he has quickly learned which tools are safe spaces for him and which are not. And I will say that, you know, it is not a secret.

It was very public that GROK was outputting significant anti Semitic content that was super harmful to my Jewish son. And putting our students in environments, whether it be Jewish students or students of color, where they are being fed biased and hateful content is not something we want in our systems. And so I wanted to ask.

We heard a little bit earlier about the ability to report that. I don't know if anybody here has had the opportunity to do so and whether there has been follow up or what your experience with that reporting mechanism has been.

Yeah, please. So I tried to create an image through. This is Microsoft Pilot. No. Yeah, this is Microsoft Pilot. So I prompted it to create an image of a Native American person in the sciences. Scientists. Right. So what it came up with was a person in a headdress with a white lab coat.

And I'm like, this is not what I want. So I mean, I was curious, I wanted to see what kind of images that came up. And that's what I got.

And I was very disappointed, especially as a parent of Filipino, African American, Native American children who are in the sciences, one in high school, one who just started his third year as a math major. So there are things like that. I'm like, okay, we'll report it, but you get an automated email, that's it.

So you know, I mean, again, got to keep humans in the center. Right. So as much as concerns that Zoe's raise, to what extent will they be addressed? Well, thank you for asking that.

Yeah, so I think. Thank you for that. Sadly, I think we could have example after example of offensive things that come out of these tools to your point. And so I guess it sounded like the CSU system is relying on the ChatGPT embedded reporting tool.

It wasn't clear to me what happens next, which is kind of why I asked if anybody on this panel had experienced it. I don't know if someone from the CSU knows the answer. It didn't seem they did. It seemed like they're relying on ChatGPT to address it direct. Is that right?

I'm looking at them, but I don't know if you want to come back on Mr. Clark, but worth following up because I do think that at least in the EDU system we are putting out as California and the CSU, I would want to know that that was being addressed.

That if something like this happened and a student was proactive enough or faculty frankly to report it, as it says, there is a mechanism that there's a change that that is corrected going forward yet. Did you announce Ming?

Yeah. I personally don't have experience with. I didn't know that you had the ability to report if ChatGPT was producing content that was inappropriate. But I will say, like in the one system wide message that was sent out to students, because I received it as well, there wasn't any referencing of that.

And I think that's something important that the system needs to share is if there are inappropriate things. Like I understand that there's probably six different websites that will lead you to a message that says, hey, if there's something wrong, you should report it and let us know.

But I think that should be kind of headlined is, you know, here's ChatGpt EDU, here's the training link, and here's a link to let us know if things aren't looking right, you know, and it's not really shared in that whole. Based off of my experience.

Yeah. And I would want that in the training too. Absolutely. People should be mandated to take, I.

Will add for staff. It's a huge concern because of discipline. They're afraid that what if I put something in and chat did something, am I going to get in trouble now because it flagged this as inappropriate and then I don't know where it flagged who it went to. Did it go to my manager?

Did it go to hr? Did it go. So that's our concern too is we don't know where this data is going because staff has been rolled out and we don't have any training. It's not required, right?

No. And that is, I think it should also be clear that to the extent that you're reporting biased, problematic outputs, that that is not a cause for discipline for anybody, students, staff, faculty, period. Because that's important to happen and it shouldn't be something that results in discipline by any means. Thank you for that.

So hopefully we can get follow up on that too. There's lots of follow up coming out of this hearing which is good news. And then, you know, there was a comment made that it was rolled out without consultation. So my understanding is that it was rolled out.

Students, faculty had access to this and then these work groups that some of you are on were convened that allowed for input after the fact. Is that correct? Am I understanding that?

Yeah, I think it speaks to what I said. Okay. As staff in the system, we don't feel that we were included. There may have been one staff person maybe on the system wide, but we don't know. I can't tell you who that person is or what their name is. The information.

We usually get one seat and it's very hard for that person to even translate out to the rest of. We're, you know, 20,000 staff.

Yeah.

Right. And so for us, it's an afterthought. And then we're the ones, we're the behind the scenes helping. When the question comes in for help, it's not going to faculty, it's going to staff. And so we have staff who are now concerned because they're like, it's here.

And we actually don't know what to do with it because we weren't asked or involved or how this is going to affect my day to day work and life. And how do I assist the faculty and staff Member that make and student that may come to the room to ask me what do I do?

And so that's what we were talking about. We weren't a part of it from the beginning. We could have helped probably roll this out to our staff in a way that would make it transparent. And we're here to help. We're not against it. Of course, we believe in technology as well.

But also at the same time, we weren't at the table.

In a meaningful way.

Yeah.

Yeah. No, I appreciate that. Does anybody else want to chime in for anything besides what's been added?

I'm good.

Okay, go for it. Yeah, I can. I can kind of speak to this. So. So I also spoke to the. I think several of us did bring this up. Faculty were initiating conversations, waving flags like, hey, we need to. We need to consider what's going on with generative AI in our classrooms and how it's being utilized.

And so. And we kind of waved a flag in 2019. We didn't see overt interaction at that point. And that's why it was. We did another call in 2023.

When the announcement came out early this year, although we'd been working, the AI comments had been up, there'd been a report from the Gen AI Committee that integration was great. What we didn't see was on the rollout with the agreement with OpenAI, we didn't see. And again to Catherine, excuse me, President. Here.

To the points made is that. The. The rollout, I think we could have done a much more in depth job in that way and we could have isolated some areas that we still had gaps.

I don't think that the idea of an AI literate CSU and someone and a CSU that's leading the nation in this work is a bad idea at all. I think it's excellent that we're trying and I want to applaud that in many ways.

But at the same time, that consultation piece is so critical to what we stand for as a CSU and how we move forward with things arm in arm in these endeavors is really, really critical. And so there has been a lot of collaboration as it's been rolling out.

But then there's also this sort of, now we also have this added burden of dealing with what didn't get taken care of before it rolled. I'll leave it at that.

No, I appreciate that. And I think, you know, I'm a little biased here because I'm a former Professor myself and I do think that the magic of these institutions are, is everyone in the classroom, that's where the magic happens, where the students grow and learn and change and the faculty is growing and learning with them.

And nobody knows these students better than the people who are in the classroom with them day in and day out, sitting in office hours, answering their questions, hearing their deepest fears. And so I do think that it is critically important that we put those critical people at every table.

And I think some of the questions we asked about what was in the contract today that it sounds like maybe aren't in the contract would be had there been more people at that table.

And so I do think it's really important, I had the privilege over interim to go and have conversations with some of the labor leaders in the United Kingdom who are dealing with AI. And it is a country that has long been labor leading in the world.

And they, I think similar to the, to a lot of what we're heard today, the application of AI is different in every single context and you cannot paint it with a broad brush. And so it is very complicated to create policies that will apply in every setting. That's just the reality of this tool.

And so the way they are facing this future is by empowering the workers, by making the workers seat at the table around AI, a loud and powerful one.

And I actually think that it's an interesting way to go because it's not saying we're going to ban AI in this setting or we're going to whatever it is, because you can't do that. AI has pretty much good applications and potentially harmful applications in every setting.

But these are the people who know the context, and I think it's the way we should be thinking about approaching this in the workforce. And something was said about replacement. I think it was you that said it.

And that was an interesting thing to hear, because what I read about this rollout, that was not something that came to mind for me. I don't think the way that this has been talked about is something that would, I guess, in my mind, ever replace the faculty. Like, that's for staff. I can't imagine that.

But that's something that we should always keep our minds on. I think AI is best used when it's augmenting our incredible powers versus replacing them. And so I appreciated that. And I think, again, putting the people at the table is how you ensure that it is powerful augmentation instead of replacing people.

And a University wouldn't be what it is without the people. So I want to thank you all for being here.

Thank you so much, Madam Chair. Just to amplify that, it's all about the people at the table. You are doing the tremendous work each and every day for our system, for our students. Really appreciate all your leadership and efforts. Just a quick comment or question on the reporting aspect.

We know that we've directed CSU to establish a civil rights office. Is this a place that maybe the reports can go to? Wanted to just put that out there for problematic content. Any thoughts on that? Or maybe to the CSU?

I think there's an opening there. Yes, absolutely. One that I would love to take a look at how that might. That might work. And I know that there's a number of faculty that are also. And students and staff that are interested in what we're doing in that space in General and certainly would add. Thank you for the.

The idea.

Thank you so much. Sounds like we could definitely follow up and chat about this. So thank you. Appreciate that.

Thank you.

Any other thoughts? All right. Any other questions? Comments? Well, thank you so much for a robust panel, and thank you so much for everything you do each and every day. And thank you for sharing your powerful insights. Appreciate it. Thank you. Thank you.

Thank you as well.

At this time now, we'd like to invite Members of the public to speak each speaker will have one minute to provide any comments. If you'd like to speak, please join us at the microphone in the center and you'll have one minute to speak.

Welcome. Thank you. Mitch Steiger with CFT, a union of educators and classified professionals. Appreciate the opportunity to testify today. Just wanted to clarify, we don't represent workers at CSU, but very much wanted to stand in solidarity with them and share a lot of the concerns that they raised about the introduction of this technology.

We hear identical concerns from our Members, both in the higher ed System and in TK12, about this technology really being forced on workers and students with little to no coordination with them or meaningful review from them or their representatives with predictably concerning results.

And we would strongly urge that not only should workers be involved in this, but it really should start and end with the perspective of workers and students, and that if this technology is not going to work for them, then maybe it shouldn't be introduced in a given context.

And there is also a real financial concern here that was also touched on by some of the panelists where we've seen a lot of schools spend a lot of money introducing a given form of technology, only to find that it didn't work and then it's withdrawn and all that money could have gone to a much more positive purpose.

So we would really urge not just centering workers and students in this, but starting with them and making sure that it works for them and that key principles like it is not going to replace workers, where some of those developing this technology are pretty open about that being the goal.

And then it's also not going to harm the critical thinking abilities of students. Some of the developers are much more interested in seeing what the technology can do than making sure that there aren't going to be harms there. When you talk to them, they say, well, that's the job of the Legislature.

It's not my job as a developer of the technology. So we would just urge the Legislature to keep that perspective in mind, make sure that workers are protected and that the needs of workers and students drive this conversation. Thank you.

Thank you so much.

Good afternoon. Thank you for convening this panel. I'm Kevin Weir, Professor of sociology at the Sacramento campus. You asked about replacement and suggested that that was imponderable. It is actually what keeps me up at night.

We had on my campus a manager who had a little bit too much time on his hands, and so he started creating different bots using AI, including a contract interpretation bot where he fed the faculty contract into a bot and then you could query it to get advice on how to interpret the contract.

We got that taken down, but before that we tried it and it gave really bad advice. This same fellow also created tutorial bots for classes that is clearly the work of graduate student assistants. So work replacement is absolutely a thing. The sine qua non of it was a Concordia University art history Professor.

You can Google this to see the news stories. His recorded lectures were continued to be used and his class was taught and he had been dead for four years. He died in 2019 and his course was taught through the pandemic. So yeah, it's the kind of thing that I worry about a lot.

Thank you so much. Any final comments, public comments Sienna thank you so much for all the public comments and very grateful to Chair Bauer Cahan and team. Do you have any closing comments?

Chair Bauer-Kahan thank you so much again to Chair Bauer-Kahan, to the Privacy Committee, to the Assembly Higher Education Committee and to each and every one of you for the thoughtful conversations here today. And thank you for all the recommendations that have come out as well.

Thank you again to our panelists for speaking with clarity and Commission on the topic. Like any new disruptive technology, AI presents both promise and peril. Users need relevant and timely training on the technology itself and leaders need to have the time and forums to transparently consider how these tools integration can work.

And today we've had that really robust discussion here today as a starting point and really a deep dive into a number of the recommendations and other areas that we can look into.

And we know the ethical use of these technologies is paramount, important so must thoughtfully consider the impact of this technology in our classrooms, in our workplaces and in higher education throughout.

We've heard a number of concerns and also about the rollout, but also looking at making sure that the safety is paramount, looking at system wide policies to ensure consistency across the system as we l.

We really look forward to the continued partnership here today with my colleagues on these committees and to all the stakeholders in this room and everyone watching online.

But thank you so much again to all the Members, to all the speakers, all the public participants who spoke with such passion and clarity today and we look forward to future conversations around this topic. Thank you so much. Specific to the to Kevin and the Assembly Higher Education Committee staff and to the Privacy Committee staff as well.

Really appreciate everyone's hard work and efforts on that and to all our colleagues as well to Chair Power K. And once again with that this Committee meeting is adjourned. Thank you thank you.