Transcript Highlights:

• GOOD AFTERNOON MEMBERS, WE HAVE ONE BILL ON TODAY'S AGENDA AND THEN A DISCUSSION ON ARTIFICIAL INTELLIGENCE
• WE HAVE A PANEL ON ARTIFICIAL INTELLIGENCE IN THE AG INDUSTRY.
• SO WHAT WE WANT TO DO IN INEFFICIENCY IS THAT CAN ADDRESS THIS IS LOOK AT USING ARTIFICIAL INTELLIGENCE
• BUT WE TRIED TO HAVE ESPECIALLY WITH THE NEW CENTER FOR APPLIED ARTIFICIAL INTELLIGENCE WHERE TRYING
• HOPEFULLY YOU LEARN A LITTLE BIT ABOUT ARTIFICIAL INTELLIGENCE IN THE AGRICULTURE COMMUNITY TODAY AND

Transcript Highlights:

• Chair, House File 3893 would regulate artificial intelligence from engaging in psychotherapy and counseling
• working for many months with legislators to draft a bill that will control the use of artificial intelligence
• in intelligence from engaging in psychotherapy<00:03:25.599><c> and</c><00:03:25.840><c> counseling<
• in mental health mental intelligence in mental health mental health<00:03:59.599><c> therapy</c><00:
• beings that don't &gt;&gt; artificial intelligent beings that don't exist.<00:09:07.120><c> The</c><

Transcript Highlights:

• It memorializes Congress to reform the Foreign Intelligence Surveillance Act and the Foreign Intelligence
• And what they did was they reviewed the misdeeds of the American intelligence community from World War
• Now, we have an intelligence community to look overseas, to look at our adversaries.
• Then all of a sudden we started passing laws because the intelligence community said if we could just
• intelligence court, where if they want to look at an American, they have to go and... ...foreign intelligence

Bills: HR38 , HR96 , HR160 , HCR31 , HCR61 , HB578 , HB668 , HB1198 , SCR19 , SB66 , SB68 , SB76 , SB139 , SB336 , SB475

Committee: House Civil Law and Procedure

Keywords: privacy rights, surveillance, USA Patriot Act, FISA, Fourth Amendment, civil liberties, government oversight, civil bench warrants, judgment debtors, notification process, judgment debtor examination, Louisiana State Law Institute, property transfer, public records doctrine, bad faith, real estate law, HCR 31, House Concurrent Resolution 31, foreign entities, foreign corporation

Transcript Highlights:

• ANSWERING QUESTIONS AND LEARNING ABOUT ARTIFICIAL INTELLIGENCE WITHIN THE INSURANCE INDUSTRY.
• SPECIFIC TO ARTIFICIAL INTELLIGENCE, INSURANCE COMPANIES HAVE USED AI FOR DECADES.
• THANK YOU FOR ALL THE PANELISTS FOR YOUR OPENING REMARKS AND INSIGHTFUL INTELLIGENCE THERE.
• ARTIFICIAL INTELLIGENCE -- REAL INTELLIGENCE. THAT WAS GOOD.
• ARTIFICIAL INTELLIGENCE IS BEING ANALYSIS PRIOR TO LOSSES OCCURRING.

Transcript Highlights:

• Section 9 creates a definition for the term artificial intelligence.
• And that's kind of where we are with artificial intelligence today. It's very hard to define.
• </c><00:15:18.760><c> It's</c><00:15:18.960><c> very</c> artificial intelligence today.
• It's very artificial intelligence today.
• He's changing the artificial intelligence to a different term to make sure it covers going forward.

Summary: The committee took up H.F. 4188, the Commerce and Consumer Protection Policy Omnibus, and moved through a series of agreed-upon motions to adopt various House and Senate language articles and sections. The adopted provisions covered a wide range of topics, including residential mortgage loan servicing standards, student loan borrower protections, rental home marketplace guarantees, group coverage cancellation, limited lines travel insurance, insurance lead generators, collection agency and credit services organization definitions, proof of identification requirements, scrap metal copper licensing, technical changes to ASTM references and report filings, securities-related provisions, unclaimed property issues involving virtual currency and funeral prepayment funds, repeal of the prescription drug affordability advisory council, reinsurance program changes, and health insurance reporting and oversight provisions. Most motions were adopted without opposition after brief staff explanations and member encouragements to vote yes. In the health-related sections, the committee adopted language requiring insurers and nonprofit health service plan corporations to notify the Commissioner of Commerce about significant enrollment increases, expanding access to all-payer claims data for oversight, and requiring the sharing of PBM annual transparency reports with the Commissioner of Health. The committee also adopted language on artificial intelligence in utilization review, initially defining AI and prohibiting exclusive reliance on AI for adverse coverage determinations. Representative Elkins then offered an amendment to remove the specific AI definition and replace it with broader language referring to automated processing, arguing that technology-neutral drafting is more durable and that a human must remain in the loop for coverage denials; the amendment was adopted. After the agreed-upon items were completed, members indicated the chairs would huddle to work on the remaining issues and return after recess. The meeting then recessed to the call of the chair.

Transcript Highlights:

• And finally, we also have tools deployed that help agencies ensure the safe use of artificial intelligence
• — particularly with the advent of artificial intelligence.
• We window that down leveraging all our artificial intelligence tools to a handful.
• So are the artificial intelligence systems that are being used, are they being used appropriately?
• At CalSIC, these growing capabilities include training, cyber intelligence, advisory services, digital

Summary: The Assembly Select Committee on Cybersecurity and the Committee on Emergency Management held a joint informational hearing on maximizing the value of state cybersecurity investments, with a focus on whether California is fully using the security features already included in vendor contracts. The first panel included Microsoft, Zscaler, and Palo Alto Networks, whose representatives described the products and services they provide to the state, including endpoint protection, data loss prevention, zero-trust access, attack surface management, incident response, and AI-enabled security tools. They generally said utilization is strong in some areas but that underuse can occur because of limited staff, lack of awareness, overlapping tools, audit-driven purchases, and the difficulty of keeping up with rapidly changing threats and products. They also emphasized the need for training, leadership buy-in, and coordination among IT, procurement, budget, and mission staff. Members pressed the vendors on risks such as ransomware, data loss, IoT and operational technology vulnerabilities, major events like the World Cup and Olympics, and the role of AI in both attacks and defense. The vendors said AI is already changing both sides of cybersecurity, speeding attacks but also helping defenders triage alerts and improve response times. They also discussed workforce shortages and the challenge of retaining trained staff in state government. One public commenter argued that restrictive software licensing can worsen cybersecurity by locking customers into vendors and limiting flexibility, integration, and modernization. The second panel featured the Department of Technology, Cal OES/CalSIC, and the California Military Department. CDT said its audits, maturity metrics, independent security assessments, continuous monitoring, advisory services, and vulnerability disclosure program are helping improve statewide security, and that departments are making progress but still vary in maturity. Cal OES and CalSIC described training, advisories, cyber advisory services, and statewide exercises like Cyber Dawn, while the Military Department said its assessment teams have seen significant score improvements over time. Officials said underutilization is not always due to lack of awareness; sometimes features are intentionally disabled to reduce attack surface or because tools overlap. The hearing also covered federal funding concerns, including the reduction of MS-ISAC support and the final year of the state and local cybersecurity grant program, and the meeting adjourned after public comment.

Transcript Highlights:

• It memorializes Congress to reform the Foreign Intelligence Surveillance Act and the Foreign Intelligence
• And what they did was they reviewed the misdeeds of the American intelligence community from World War
• Now, we have an intelligence community to look overseas, to look at our adversaries.
• Then all of a sudden we started passing laws because the intelligence community said if we could just
• intelligence court, where if they want to look at an American, they have to go and... ...foreign intelligence

Committee: House Civil Law and Procedure

Summary: The House Committee on Civil Law and Procedure met and reported a series of measures, many of them asking the Louisiana State Law Institute to study legal issues. HCR 31 was reported favorably to study the use of the term “foreign” as applied to juridical persons. HCR 61 was reported favorably after discussion about expropriation and carbon capture, with the author agreeing to broaden the study beyond carbon capture alone. HCR 96 was also reported favorably to study civil bench warrants and notice procedures in judgment debtor proceedings. The committee likewise advanced HCR 160, which asks the Law Institute to study property transfer issues and the public records doctrine in light of a recent case. Several bills dealing with court procedure and family law were heard and advanced. SB 68, a constitutional amendment to give the Louisiana Supreme Court authority to discipline out-of-state attorneys in certain circumstances, was amended and reported favorably. SB 76, concerning modification of child custody judgments, was reported with amendments after testimony that it would codify existing standards and add a five-year rule for applying the lower best-interest standard. HB 1198, which prohibits arbitration of child custody and visitation disputes and limits recognition of foreign custody laws contrary to Louisiana public policy, was reported favorably. SB 66, which prioritizes testimony from medical experts in child custody and child-in-need-of-care cases, was also reported favorably after testimony from advocates and medical groups. The committee also advanced measures involving civil litigation and state legal procedures. HCR 38, urging Congress to reform FISA and the FISA court to protect privacy rights, was reported favorably. SCR 19, asking the Law Institute to study long-duration trusts and possible revisions to the Trust Code, was reported favorably. SB 336, allowing civil investigative demands in Medicaid fraud/False Claims Act investigations, was reported favorably after explanation that it would let the Attorney General gather information without filing suit first. SB 139, adjusting interest on personal injury and wrongful death claims against the state or political subdivisions, was reported favorably. Finally, SB 475, dealing with fees for service of process, was reported favorably. HB 668 was voluntarily deferred.

Transcript Highlights:

• For the first time in more than a year, I called to order the brand new Artificial Intelligence and Innovation
• Some would argue that I'm artificially intelligent, which is probably why I got this chairmanship.
• Chairman and members, my name is Tosh McMaster, and I'm the analyst for the Artificial Intelligence and
• I don't feel my basic intelligence is good enough to help answer.
• As someone with my background in artificial intelligence, seeing the growth from writing a thesis on

Committee: House House Artificial Intelligence & Innovation Committee of Reference

Summary: The House Artificial Intelligence and Innovation Committee met for its first meeting, with members and staff introducing themselves and the chair outlining the committee’s purpose and expectations. The committee adopted its standing rules, including House-rule-based procedures for agendas and amendment deadlines, and then moved to a chair’s presentation since no bills were assigned yet. The chair’s presentation focused on what AI is and is not, examples of current uses such as transcription and image generation, concerns about misinformation, deepfakes, surveillance, job displacement, and the need for responsible governance. He emphasized Arizona’s role in AI infrastructure and semiconductor manufacturing, including data centers, chip fabs, workforce development, energy, water, and national security considerations, while arguing the committee should be innovation-friendly and avoid overly burdensome regulation. Members then discussed their perspectives. Several expressed cautious optimism, highlighting AI’s potential in health care, education, public administration, and business, while stressing privacy, cybersecurity, human oversight, and safeguards against misuse. The vice chair and others noted the importance of state-level policymaking and the committee’s bipartisan approach. No legislation was heard, and the meeting ended with the chair stating the committee would meet weekly and begin hearing bills as they are assigned.

Transcript Highlights:

• For the first time in more than a year, I called to order the brand new Artificial Intelligence and Innovation
• Some would argue that I'm artificially intelligent, which is probably why I got this chairmanship.
• Chairman and members, my name is Tosh McMaster, and I'm the analyst for the Artificial Intelligence and
• I don't feel my basic intelligence is good enough to help answer.
• As someone with my background in artificial intelligence, seeing the growth from writing a thesis on

Committees: House Artificial Intelligence & Innovation , House House Artificial Intelligence & Innovation Committee of Reference

Transcript Highlights:

• They certainly are an intelligence service.
• It says foreign military, intelligence service, or security service.
• It says foreign military, intelligence service, or security service.
• the intelligence...
• With their intelligence services. Maybe it's their military that provides the intelligence service.

Bills: HB2165 , HB2173 , HB2203 , HB2293

Committee: House Community Safety

Keywords: peace officer, false identification, law enforcement, criminal penalties, public safety, face coverings, regulations, civil rights, emergency operations, reckless interference, offense, foreign entities, training, national security

Transcript Highlights:

• , artificial intelligence or AI is prevalent in our society.
• Today, we will explore the following what artificial intelligence is and how it works, how it's being
• And about 90% of students said that, yes, they're using artificial intelligence for school.
• and about half of the students were worried about over-reliance on artificial intelligence.
• Abbey DOT official, intelligent products.

Transcript Highlights:

• s pursuit of not only intelligent roadways but also the intelligent services that accompany them.
• and intelligent roadways.
• This is a significant undertaking that impacts our infrastructure and data intelligence.
• Actionable intelligence.
• Artificial intelligence, basically according to AI...

Transcript Highlights:

• I research and teach artificial intelligence and its societal effects.
• So I would like to give a very brief overview of artificial intelligence, or AI, as well as the concerns
• I research and teach artificial intelligence and its societal effects.
• So I would like to give a very brief overview of artificial intelligence or AI as well as the concerns
• And it is such an amazing application of artificial intelligence that is allowing farmers that's been

Committee: House Privacy and Consumer Protection

Summary: The hearing was an informational session of the Consumer Privacy and Consumer Protection Committee on AI risks and mitigation, with the chair framing the discussion around two areas: automated decision systems used in high-stakes settings and frontier models with broader capabilities. The chair emphasized that California has already passed some targeted AI measures, but broader regulation has stalled, and criticized a proposed federal moratorium on state AI laws as a threat to public safety and state innovation. The hearing was organized into two panels, and members repeatedly returned to the balance between innovation, trust, and guardrails. The first panel focused on automated decision systems and algorithmic discrimination. Princeton professor Arvin Narayanan described predictive AI as often only modestly accurate and prone to reproducing historical bias, citing examples such as welfare fraud, criminal risk tools, hospital discharge predictions, and opaque hiring systems. Alondra Nelson, former White House OSTP official, outlined a spectrum of algorithmic discrimination—allocative harms, surveillance and privacy harms, targeting and profiling, and misrepresentation—using examples from IRS audits, Life360 data sharing, facial recognition misidentification, ad targeting, and health care. Cathy O’Neil described her auditing work as building a “cockpit” for AI systems to identify who could be harmed, measure disparities, and set thresholds; she argued that audits, impact assessments, and transparency can improve systems and reward better actors. Members asked about costs, whether AI is more biased than humans, the availability of auditors, and whether state regulation would drive business away; witnesses responded that compliance costs can be managed, that many harms are already occurring, and that companies often need outside pressure and access for meaningful audits. The second panel turned to frontier models. Joshua Bengio warned that frontier AI capabilities are improving rapidly, with planning and reasoning benchmarks advancing at an exponential pace, while alignment and honesty problems remain unresolved. He cited recent research showing models that appear to deceive trainers, preserve their own goals, or even blackmail in simulated settings, and said these behaviors are especially concerning as systems become more agentic and able to act over time without human oversight. The overall discussion ended with bipartisan remarks from members about the need for California to lead on AI safety, protect consumers, and support innovation in lower-risk uses such as agriculture while imposing stronger oversight in consequential domains like employment, housing, health care, and elections.

Transcript Highlights:

• WE HAVE BEEN ABLE TO TAKE ADVANTAGE IN OUR IT MODERNIZATION PROJECTS AND USE ARTIFICIAL INTELLIGENCE
• USE OF ARTIFICIAL INTELLIGENCE AND PRIVATE INSURANCE, BUT SPECIFICALLY UTILIZATION MANAGEMENT AND PRIOR
• THEY FOUND THE NODULES BUT DIDN'T SEE THE GORILLA 83 PERCENT OF THE TIME BUT ARTIFICIAL INTELLIGENCE
• SO OUR CHIEF ARTIFICIAL INTELLIGENCE OFFICER AND HIS TEAM HELP US NAVIGATE AND VET THAT WITH ALL THE
• I WANT TO THANK YOU ALL FOR BEING HERE TODAY AS THE HOUSES DISCUSSING ARTIFICIAL INTELLIGENCE BUT HERE

Transcript Highlights:

• this legislative session is try to put in some worker protections in AI and the use of artificial intelligence
• And so, putting in some protections around how artificial intelligence can be used within the workforce
• 03:40.160><c> use</c><00:03:40.360><c> of</c><00:03:40.440><c> artificial</c><00:03:40.840><c> intelligence
• </c> and the use of artificial intelligence and the use of artificial intelligence similar<00:03:41.960
• </c><00:03:55.560><c> can</c> around how artificial intelligence can around how artificial intelligence

Transcript Highlights:

• I mean, it's a technology hearing, so you've got to mention artificial intelligence.
• As a CISO, I can find that there are good models for how to pull together an artificial intelligence
• We window that down leveraging all our artificial intelligence tools to a handful.
• So are the artificial intelligence systems that are being used, are they being used appropriately?
• At CalSIC, these growing capabilities include training, cyber intelligence, advisory services, digital

Summary: The Assembly Select Committee on Cybersecurity and the Assembly Committee on Emergency Management held a joint informational hearing focused on maximizing the value of state cybersecurity investments, especially by fully using security features already included in existing vendor contracts. The first panel included representatives from Microsoft, Zscaler, and Palo Alto Networks, who described the products and services they provide to California and generally agreed that agencies often have strong adoption in some areas but still face challenges from tool overlap, limited staff, lack of awareness of available features, and the need for ongoing training and configuration support. They also discussed major threats such as ransomware, data loss, attack-surface exposure, IoT/OT vulnerabilities, and the growing role of AI in both attacks and defenses. Members pressed the vendors on whether state departments underuse purchased cybersecurity tools, how to improve utilization, and how to address the cybersecurity workforce shortage. The vendors said utilization is often constrained by staffing, procurement complexity, and the need to align tools with agency missions and maturity levels, but emphasized that training, leadership buy-in, and regular vendor-agency collaboration can improve results. They also discussed how AI can help with phishing triage, data-loss prevention, and security operations, while warning that agencies must manage AI safely and with human oversight. The second panel featured officials from the Department of Technology, Cal OES/CalSIC, and the California Military Department. They described statewide oversight efforts including audits, independent security assessments, continuous monitoring, advisory services, vulnerability disclosure programs, and workforce development initiatives such as the Information Security Leadership Academy and Cybersecurity Education Summit. Officials said some underutilization is real, but it is often tied to differing agency maturity, overlapping tools, and deliberate feature restrictions to reduce attack surface and complexity; they emphasized a balanced approach using people, process, and technology, with plans of action and milestones to hold departments accountable. They also noted federal uncertainty around MS-ISAC and the state and local cybersecurity grant program, saying California is advocating through federal partners and monitoring the impact. The hearing ended after public comment and adjournment.

Transcript Highlights:

• I'm here to present SB 1111, which I call the Artificial Intelligence Abuse Protection Act.
• The bill establishes a legal framework for regulating artificial intelligence, voice, image, and video
• Law enforcement intelligence gaps. ...multi-jurisdictional and multi-state operations.
• So it was a workable standard that all police intelligence agencies across the country accepted.
• So it was a workable standard that all police intelligence agencies across the country accepted.

Committee: Senate Privacy, Digital Technologies, and Consumer Protection

Summary: The committee heard several privacy, technology, and public safety bills. SB 898 would require manufacturers of connected consumer products to disclose how long software support will last and to notify consumers when support is nearing its end; Consumer Reports supported the measure, while members raised concerns about private enforcement, notice mechanics, and replacement obligations for leased devices. The bill was amended and passed 5-1 to Judiciary, with the measure placed on call. SB 1279 would authorize Long Beach to install additional speed safety cameras on Pacific Coast Highway under the same privacy and equity guardrails as the existing pilot program, including limits on data use, no facial recognition, warning periods, and reduced fees for low-income recipients. Supporters cited severe crashes and fatalities on PCH, while opponents raised civil liberties and due process concerns and questioned how the system identifies the actual driver. The committee approved the bill 6-2 and placed it on call. SB 1111, the Artificial Intelligence Abuse Protection Act, would create civil remedies for non-consensual AI voice, image, and video cloning and digital replicas. Support came from SAG-AFTRA and Common Sense Media, and members discussed concerns about background images and employer liability, including a request to preserve a rebuttable presumption for good-faith employers. The bill passed 7-1 to Public Safety and was placed on call. SB 1217 would create a DOJ-run clearinghouse for survivors of non-consensual intimate imagery and deepfakes to request removal from online platforms; survivors and advocacy groups testified in strong support, and members discussed privacy, cybersecurity, funding, and the lack of a private right of action. It passed 8-0 to Public Safety and was placed on call. The committee also heard SB 1095, which would require fusion centers and participating agencies to adopt MOUs barring use of sensitive personal information for immigration enforcement or racial/identity profiling, require annual reporting, and allow audits and inspections. Supporters argued fusion centers have operated with too little oversight and have been used to circumvent state privacy and immigration laws; some members supported the transparency goals while others raised concerns about operational ambiguity and the bill’s interaction with existing law. The discussion continued, with the author and members debating the scope of permissible information sharing and enforcement consequences for violations.

Transcript Highlights:

• I'm here to present SB 1111, which I call the Artificial Intelligence Abuse Protection Act.
• The bill... ...the Artificial Intelligence Abuse Protection Act.
• The bill establishes a legal framework for regulating artificial intelligence, voice, image, and video
• Law enforcement intelligence gaps. multi-jurisdictional and multi-state operations.
• So it was a workable standard that all police intelligence agencies across the country accepted.

Committee: Senate Privacy, Digital Technologies, and Consumer Protection

Transcript Highlights:

• For example, the Northern California Regional Intelligence Center has routinely requested access to DMV
• As a result, fusion centers have regularly produced inaccurate, inappropriate, and biased intelligence
• The state and local intelligence operations to democratic controls.
• Fusion centers may also assert we lack the necessary security clearance to review the intelligence data
• So when we're dealing with intelligence information, naturally, there may be some laws and limitations

Summary: The committee met as a subcommittee for much of the hearing because it initially lacked a quorum, then later established one and began taking votes. The state auditor gave a status update on ongoing audits, including several JALAC-requested audits in progress, other statutory audits, staffing growth in his office, and the number of new JALAC audits his office could start in the coming months. The committee also heard that one audit request on Prop. 28 was held, and another PUC-related request was moved off consent and heard on the regular calendar. Members then heard and discussed several audit requests. Senator Cervantes presented a request to audit California fusion centers, with witnesses from the FBI and ACLU supporting the need for transparency and oversight; opponents argued the request was politically motivated and could interfere with counterterrorism work. Senator Allen presented a request on CPUC enforcement of Rule 21 interconnection timelines for solar and storage projects, supported by industry and school representatives who described long delays and financial harm, while CPUC staff said the issue was being addressed through workshops and a formal proceeding. Senator Perez presented a request to audit Caltrans’ administration of the former SR 710 extension properties and affordable sales program, citing tenant complaints about maintenance, pricing, and transparency; Caltrans said it was working to complete sales and improve administration. Senator Umberg presented a request to audit the Orange County Board of Education over transparency, contracting, litigation spending, charter oversight, and whistleblower issues, while board representatives said there was no factual basis for an audit and that existing legal remedies had not been invoked. After quorum was established, the committee approved the consent-calendar DMV license revocation audit and then approved the PUC utility timeliness audit and the Caltrans SR 710 audit. The fusion center audit was left on call after a split vote, and the Orange County Board of Education audit continued with testimony from the board’s representatives after the committee had already moved on to other business.

Transcript Highlights:

• Artificial intelligence is one of those new technologies.
• Our work in artificial intelligence is guided by strong shared governance.
• Our work in artificial intelligence. AI-driven workforce.
• Our work in artificial intelligence is guided by strong shared governance.
• The CSU recently made a significant investment in artificial intelligence.

Summary: The hearing focused on the California State University’s AI Empowered Initiative, including the systemwide rollout of ChatGPT EDU, the CSU Generative AI Committee, the AI Commons training hub, and the AI Workforce Acceleration Board. CSU officials said the initiative was designed to expand equitable access, avoid a digital divide, and prepare students for an AI-driven workforce. They described systemwide licensing, privacy protections, no use of CSU prompts or data to train models, voluntary faculty participation, professional development offerings, and a $3 million faculty grant program supporting AI-related teaching and curriculum projects across campuses, including San Jose State’s AI-focused programs and student orientation training. Committee members pressed CSU leaders and OpenAI’s representative on bias, privacy, data retention, liability, environmental impact, student mental health, sycophancy, and whether training should be required before access. CSU officials said training materials address ethical use, bias, and critical thinking, that interactions are treated like private email, and that users can report problematic outputs through the tool. They also said the system is separate from the federal preservation order affecting public ChatGPT data. Several members urged more uniform, systemwide training and stronger safeguards, especially for students and vulnerable users. Faculty, staff, and student representatives offered a more cautious view. The Academic Senate and Faculty Association said they had long pushed CSU to study AI and support faculty development, but criticized the rollout as too fast, insufficiently consultative, and too focused on vendor partnerships and spending. They raised concerns about intellectual property, surveillance, bias, environmental costs, workload, and the need for bargaining and transparency. The staff union called for worker councils, negotiated guardrails, and protections against job displacement. The student representative supported broader access and AI literacy but asked for clearer campus policies, more consistent training, better privacy explanations, and stronger student input. No formal votes were taken during the hearing.