Transcript Highlights:

• Do you use algorithms to aggregate data and and see the outliers, what do algorithms do for you?
• Right, algorithms don't, and I believe in statute we... you know, after the algorithm, you begin investigation
• at a collection and seeing and helping with the. algorithms to find those outliers?
• I'm also the chair of Feeding Texas at the moment.
• That's when the algorithm catches you. you.

Keywords: 1185, senate, all

Transcript Highlights:

• AI can't do anything on its own, and no AI algorithm is able to make something out of nothing.
• It may be that data set has no Taylor Swift in it, and the algorithm created a song that sounds really
• They have the data, they have the algorithm, and the infrastructure.
• That algorithm can then start to try to create some kind of profile of your identity.
• I'm just dying to know: what did you feed those kids?

Summary: The Assembly Privacy and Consumer Protection Committee met with a new membership roster and adopted its committee rules after quorum was established. The hearing then began with AB 412, the AI Copyright Transparency Act, which would require generative AI developers to provide copyright holders notice when registered copyrighted works are used in training data. The author and supporters, including SAG-AFTRA, the Transparency Coalition, voice actors, writers, labor groups, and other creators, argued the bill would give artists a practical way to learn whether their works were used and to vindicate their rights. Opponents, including EFF, CalChamber, RIAA, CCIA, Chamber of Progress, Bay Area Council, BSA, and TechNet, said the proposal was technically unworkable, could burden startups, conflict with existing law and pending litigation, and raise federal preemption concerns. Members discussed the bill’s amendments, including a fingerprinting approach and narrowing the bill to model developers, and the committee voted 8-2 to pass AB 412 as amended to the Judiciary Committee. The committee then heard AB 446, which would prohibit “surveillance pricing,” or the use of personal data to charge different prices for the same product or service. The author and supporters, including Consumer Watchdog, UFCW, labor organizations, and consumer/privacy groups, described examples of differential pricing tied to device type, location, shopping behavior, and digital price tags, and argued the bill would protect consumers from discriminatory and predatory pricing. Business and industry opponents, including CalChamber, grocers, retailers, travel, broadband, and other associations, said the bill could conflict with the CCPA, interfere with loyalty and rewards programs, and create confusion about personalized discounts and dynamic pricing. The discussion focused on how the bill would treat loyalty programs, whether existing privacy law already covers the issue, and whether the proposal would unintentionally affect legitimate discounts and promotions.

Transcript Highlights:

• Another one is can you feed in all of that code and structure and rule?
• One system may have a machine learning algorithm and a deep learning algorithm to help enable or improve
• That is a machine learning algorithm that has been trained to identify smoke.
• That could be a different algorithm or a different component of that technology.
• The first kind of smoke and image detection may be focused more on a deep learning algorithm.

Summary: The committee held a work session focused on technology in government, AI, broadband, and digital equity. Seattle CTO Rob Lloyd described the city’s AI strategy, emphasizing responsible use, privacy, security, community input, and data strategy. He said Seattle is using small pilots and partnerships to test AI for tasks such as public records processing, infrastructure inspection, and permitting, while keeping humans as the final decision-makers. Members asked about bias, liability, training on best practices, labor involvement, and public records; Lloyd said AI should remain an assistant tool, not a replacement for human judgment, and that Seattle is still testing solutions for records requests and permitting. WATech CTO Nick Stow and Deputy Director Mark Quimby discussed the state’s broader AI policy, the generative AI executive order, a sandbox with more than 15 agencies, and use cases including a resident portal, cybersecurity, and wildfire detection. They stressed consent, closed systems, human-centered design, and the need to govern all forms of AI, not just generative AI. Committee members raised concerns about federal data access, labor issues, and wildfire detection effectiveness. Spokane County IT staff described a more restrictive approach to AI, citing privacy, bias, and cyber risks. The county standardized on Microsoft Copilot as its only approved chat-style AI tool, blocked other AI chat platforms, and requires human review of all AI-generated content. They said AI is also being used by criminals for phishing and deepfakes, underscoring the need for strong policy and security controls. The committee also received an update from the Attorney General’s AI Task Force. Yuki Ishizuka said the task force has 19 members and eight subcommittees covering ethics, consumer protection, labor, health care, public safety, education, government efficiency, cybersecurity, and industry/energy. The task force is working toward an interim report due December 1, 2025 and a final report due July 1, 2026, and is reviewing recommendations through public forums and advisory committees. Ishizuka warned that a federal budget reconciliation provision could bar states from enforcing AI regulations for 10 years, and several members voiced support for state authority and asked about possible 2026 legislation. The committee then heard updates on broadband and digital equity. Commerce’s Dave Pringle said the BEAD broadband program is being reshaped by new NTIA guidance, which removed or reduced emphasis on several prior priorities, and Washington is now working under an expedited process to submit its state application by September 4. He noted that no projects have been built yet, that four counties did not receive applications in round two, and that the state is trying to keep applicants engaged through office hours and a shortened review window. The Office of Equity and the Digital Equity Forum reported increased participation, new members, and ongoing outreach to tribal, rural, and underserved communities, while previewing recommendations such as creating an interdepartmental digital equity team and improving data use. Finally, Lumen’s Robert Thoms described private-sector broadband deployment challenges, including permitting, regulation, and the economics of extending fiber, while noting continued investment in overbuild projects, a $30 low-cost service option tied to the former ACP, and work with the state, tribes, and libraries. No votes were taken; the meeting consisted of presentations and member questions.

Transcript Highlights:

• <00:03:38.720> on<00:03:39.040> your<00:03:39.280> Instagram<00:03:39.680> feed
• then it pops up on your Instagram feed. then it pops up on your Instagram feed.
• At the same time, the effects of these algorithms are increasing costs for our communities.
• At the same time, the effects of these<00:08:04.879> algorithms<00:08:05.520> are<00:08
• :05.680> increasing<00:08:06.240> costs these algorithms are increasing costs these algorithms

Keywords: 919, house, all

Summary: The committee took up House File 3117, which would impose an excise tax on social media companies based on Minnesota monthly users and data-mining activity, and adopted an A1 amendment that added clarifying language identifying social media platforms. Chair Gomez described the bill as a way to tax companies profiting from data mining and social media use, citing concerns about child bullying, misinformation, and wealth concentration. The bill was laid over for possible inclusion in the 2025 taxes bill. Supporters testified that the measure would help raise revenue from a highly profitable industry and better align the tax code with the social costs of data collection and social media use. Pastor Julie Thompson, MAPE representative Tanner Fritzinger, Council Member Sue Bud, and Eric Bernstein of We Make Minnesota all backed the bill, arguing that social media companies extract value from users’ data, contribute to mental health and social harms, and should pay more toward public needs. Bernstein also framed the tax as a way to broaden the tax base and fund schools and other services. Opponents warned that the bill could sweep in local broadcasters, newspapers, and other businesses that use digital platforms and collect some user data, and that costs would likely be passed on to consumers. Wendy Pollson of the Minnesota Broadcasters Association said the definitions were too broad and could unintentionally include local media. Deb Peters, speaking for Americans for Digital Opportunity, argued the tax would raise costs for small businesses and consumers, create legal risks, and amount to double taxation. Several members echoed concerns about regressivity, administration, and whether the bill actually addresses online bullying or data privacy, while supporters said it is a first step toward taxing a new, lightly taxed industry.

Transcript Highlights:

• They're vulnerable to online bullying and harmful algorithms that maximize social engagement but not
• The average teenager is spending five hours a day on social media, almost always scrolling through feeds
• generated by an algorithm designed to keep them glued to their phones.
• I also want to make it clear that it's not just adults that recognize how harmful these addictive feeds
• All thanks to an algorithm designed to cater to her supposed interests.

Keywords: 995, all

Summary: The House first recognized several student athletes and teams, including East Bridgewater cheerleaders, Medway athletes and teams, and the Housatonic Valley basketball team, and adopted a resolution congratulating the William Diamond Jr. Fife and Drum Corps on its 25th anniversary. Members also suspended Joint Rule 12 to advance a petition concerning an extension of a lease authorization for conservation land. The chamber then took up Senate 2581, as amended, a major education and youth technology bill. The bill was rewritten to promote safe technology use and distraction-free education, including restrictions on personal electronic devices in public schools and limits on social media access for children. Members debated whether the policy should be bell-to-bell or allow more local flexibility, and several amendments were offered. An amendment from Attleboro to preserve district-specific policies was rejected, while a consolidated amendment tightening social media restrictions and related protections was adopted. The House also considered and passed several other bills, including measures amending the Sandwich charter, authorizing Springfield to help certain residents with local tax obligations, authorizing charity alcohol auctions, and directing the Boston Police Department to waive the maximum age requirement for a specific police applicant. In addition, the House rejected concurrence on a Senate version of a higher education infrastructure bill and instead appointed a committee of conference. Later, the House accepted a conference report on modernizing cannabis laws after debate focused on regulatory streamlining, equity, ownership caps, and hemp regulation. The report was adopted by roll call, and the House then passed the amended education bill to be engrossed before adjourning to the next day.

Transcript Highlights:

• Um it needs<00:42:42.800> to<00:42:43.200> feed<00:42:43.520> off<00:42:43.680><
• c> of<00:42:43.920> data<00:42:44.880> uh<00:42:45.040> quite needs to feed
• So first is algorithmic management, then electronic monitoring and human in the loop.
• > electronic algorithmic management, then electronic algorithmic management, then electronic monitoring
• <01:21:03.760> that<01:21:04.000> are feedback on algorithms that are feedback on algorithms

Bills: HF1316, HF4369

Keywords: HF1316, child support, new hire reporting, centralized work reporting system, independent contractors, payors, employers, payor of funds, withholding orders, income withholding, child support enforcement, Minnesota Department of Children, Youth, and Families, gig economy, rideshare drivers, delivery drivers, 1099, W-9, W-4, newly hired workers, rehired employees

Transcript Highlights:

• Then they put this data into an algorithm, and that data is the algorithm that is deciding the factor
• algorithm algorithm and<02:03:09.360> that<02:03:09.599> data<02:03:10.080> is<
• consumer external data and algorithms consumer external data and algorithms and<02:18:46.000>
• <02:24:13.280> where price or wage setting algorithm where price or wage setting algorithm
• and algorithmic development and whatnot. and algorithmic development and whatnot.

Keywords: 981, all

Summary: The Senate convened with a quorum, approved the previous day’s journal, and received several committee reports and conference committee permissions. It also adopted a Senate resolution recognizing the 125th anniversary of Western Colorado University (SR 8) by a 35-0 vote, with remarks highlighting the university’s history, workforce role, mountain rescue team, athletics, and ties to Gunnison and the Western Slope. The resolution was adopted and co-sponsors were added. The chamber then took up third reading and final passage on several bills. Senate Bill 152, concerning automated vehicle identification systems, was amended on the floor with a technical correction and passed 32-3. House Bill 1322, creating civil actions related to conversion therapy efforts, prompted emotional testimony from Senator Mullica about his mother and passed 23-12. House Bill 1123, addressing sexual abuse prevention in jails, passed 27-8 after a technical amendment. House Bill 1226, reducing emissions from certain electric generating units, passed 22-13. House Bill 1076, modifying transportation statutes, passed 28-0, and House Bill 1207, requiring employer disclosure of demographic workforce data, passed 23-12. Senate Bill 115, a postconviction relief measure for certain older incarcerated people, was amended and debated at length. Supporters described it as narrowly tailored to a small number of eligible inmates and emphasized victim-rights protections and exclusions for serious offenses; opponents remained opposed. The bill ultimately passed 18-17. The Senate then approved a Committee of the Whole report advancing Senate Bill 181 and several House bills—1338, 1344, 1336 as amended, 1227, 1314, 1028, 1282, and 1317 as amended—to third reading and final passage, and later adopted the consent-calendar bills in special orders without objection.

Transcript Highlights:

• And we shouldn't allow unregulated algorithms to dictate the mental health of an entire generation. if
• The algorithm keeps feeding me videos to keep me on, which benefits the social media companies at the
• The<00:34:59.600> algorithm<00:35:00.080> keeps<00:35:00.320> feeding<00:35:00.600
• > me<00:35:00.720> videos<00:35:01.240> to The algorithm keeps feeding me videos
• to The algorithm keeps feeding me videos to keep<00:35:01.520> me<00:35:01.680> on,<00

Keywords: 1187, senate, all

Transcript Highlights:

• So, for example, in an algorithm class, they might develop an algorithm, a classical algorithm, and then
• they use AI to test that algorithm.
• Think of an algorithm, for example, as a calculator. How do you test if your calculator works?
• Robotics as well: students use and develop their own algorithms and AI software that power small-wheeled
• how it is tied to the percentage of completion for high school and beyond plan components that are feeding

Summary: The committee spent much of the meeting hearing presentations on student mental health supports at Central Washington University, Washington State University, the University of Washington, and the State Board of Community and Technical Colleges. CWU described high levels of student distress, loneliness, anxiety, depression, suicidal ideation, and basic-needs insecurity, along with a campus-wide holistic well-being model that includes behavioral intervention teams, emergency aid, peer outreach, community partnerships, and a collegiate recovery community. WSU emphasized loneliness, common presenting concerns such as anxiety, depression, PTSD, and relationship distress, and the role of housing and residence life, resident advisors, living-learning communities, and coordinated crisis response in identifying and supporting students. UW highlighted prevention and support programs through LiveWell, including alcohol and other drug consultations, confidential advocacy, suicide intervention, student needs navigation, peer health educators, and peer wellness coaches. The community and technical college system reported persistent access barriers, especially for students ages 18 to 24 and those facing housing or food insecurity, and said the legislature-funded mental health pilot at four colleges expanded counseling access, telehealth, and referral pathways, while broader system efforts focus on awareness, telehealth, basic-needs supports, and workforce training in behavioral health fields. Members asked each panel how they gather student feedback, and the institutions said they use surveys, evaluations, and ongoing outreach to students in services. Questions also focused on substance-use and recovery resources, Greek life outreach, and whether colleges have enough licensed mental health staff versus academic advisors; the community college board said not all colleges have licensed mental health providers on staff, and some rely on community contracts or telehealth. The committee also discussed whether counseling services are increasing because of post-pandemic effects or because students are more willing to seek help, with presenters saying both factors likely play a role. The committee then shifted to artificial intelligence in higher education. Washington State University, the University of Washington, Western Washington University, and the community and technical college system described campus AI task forces, governance structures, and efforts to set policies for students, faculty, and staff. Presenters said institutions are requiring course-level AI use statements, promoting ethical and transparent use, expanding access to approved tools such as Microsoft Copilot, and using AI for teaching, research, advising, and administrative support while guarding privacy, equity, and academic integrity. Members raised concerns about deepfakes, bias, student monitoring, and whether AI should be used to screen applications or evaluate student work; presenters said human review remains essential and that some institutions prohibit AI use in hiring or admissions screening. The final presentation, from OSPI, described the statewide rollout of the School Links high school and beyond plan platform under 2023 legislation, saying it will standardize career and college planning across K-12, provide better data and student guidance, and connect students to postsecondary pathways and employers; OSPI said the rollout is underway but current funding only extends through June 30, 2026.

Transcript Highlights:

• on the bottom left here is algorithmic on the bottom left here is algorithmic and<01:08:38.560><
• <01:19:45.040> um to you in an algorithmic situation. um to you in an algorithmic situation
• <01:19:52.800> is on those incentives the algorithm is on those incentives the algorithm is
• <01:20:33.920> to protocol or trust this algorithm to protocol or trust this algorithm to
• We did go over algorithmic stablecoins.

Keywords: 1189, house, all

Summary: The meeting began with roll call and introductions of commission members and guests, followed by approval of the agenda and a motion to approve the February 10 minutes with a correction clarifying that one quoted statement was misattributed. The commission then moved into presentations. The main presentation came from the Conference of State Bank Supervisors on implementation of the federal GENIUS Act for stablecoins. The speaker reviewed the OCC’s recent 367-page proposed rule, noting it raises many open questions and design choices for states, and discussed expected upcoming rulemaking from the FDIC, Federal Reserve, and Treasury. The presentation focused on six areas: permissible issuer activities, reserve assets and redemption, risk management and supervision, treatment of state-qualified issuers, capital/operational backstops, and foreign issuers. It also flagged unresolved issues around Bank Secrecy Act/AML requirements and the meaning of “digital asset service provider” activities. A substantial portion of the discussion addressed yield restrictions, with the presenter explaining the OCC’s broad definition of yield and its rebuttable presumption against issuer-affiliated or related third-party yield arrangements. The speaker said this likely forecloses many existing white-label structures but leaves some room for third-party payments depending on distance from the issuer, and noted ongoing Senate debate over similar provisions. The presentation also covered reserve valuation, liquidity and diversification requirements, redemption timing, and supervisory expectations such as third-party oversight, IT security, exam cycles, and reporting. No additional votes or formal actions were taken beyond approving the amended minutes.

Transcript Highlights:

• of rent-fixing algorithms.
• So we also should not allow it to happen in algorithms.
• Algorithms are no longer just tools. They are decision makers.
• Algorithms are no longer just tools. They are decision makers.
• Algorithms are no longer just tools. They are decision makers.

Summary: The committee met as a subcommittee without quorum at first, then later established quorum and continued hearing several bills. SB 27, the annual CARE Court cleanup bill, would require courts to consider CARE referral for certain misdemeanor defendants found incompetent to stand trial, combine some hearings, allow limited data sharing among licensed medical professionals, and expand eligibility to include mood disorders with psychotic features. Supporters said it would clarify the meaning of “clinically stabilized,” streamline the process, and help more severely ill people receive treatment; opponents argued it would expand CARE Court too broadly, strain county resources, and divert attention from housing and voluntary services. The bill passed to the Health Committee on a roll call vote and was placed on call. SB 82, dealing with so-called “infinite arbitration clauses,” would limit consumer contract arbitration provisions to disputes arising from the product or service actually purchased. The author and supporters said the bill would stop companies from forcing arbitration in unrelated claims and would not ban arbitration itself. Opponents from business and banking groups argued the language was too restrictive, could create litigation over related transactions, and should be clarified as prospective only. The committee approved the bill and placed it on call after a roll call vote. The committee then heard two reparations-related bills. SB 437 would direct CSU to develop a genealogical methodology and framework for verifying descendants of enslaved people, with oversight, reporting, and guardrails tied to recently allocated state funding. Supporters said the bill would create a fair, evidence-based process; opponents, including professional genealogists and reparations advocates, argued the work is already well understood, the bill is unnecessary, and it could delay action. SB 518 would create a Bureau for Descendants of American Slavery within state government, with divisions for genealogy, property reclamation, outreach, and legal affairs. Supporters framed it as needed infrastructure to implement reparations recommendations; opponents objected to locating it in the Department of Justice, warned about data privacy and law enforcement control, and criticized the inclusion of broader communities. Both bills were moved to Appropriations and placed on call. The committee also heard SB 52, the End AI Rent Hikes Act, which would prohibit the use of algorithms to collude on and artificially inflate rental prices; the author and supporters described it as a response to AI-assisted rent fixing in California’s housing market.

Transcript Highlights:

• On top of that, it prohibits algorithmic boosting to kids under 16, which prevents big tech from pushing

Bills: SB161, SB195, SB216, SB245, SB246, SB257, SB258, SB260, SB278, SB281, SB283, SB93, SB98, SB99, SB306, SB314, SB315

Keywords: vehicle safety, crash test, crashworthiness, NHTSA, National Highway Traffic Safety Administration, New Car Assessment Program, NCAP, motor vehicle safety standards, 49 CFR parts 571 and 572, THOR dummy, THOR-50M, THOR-05F, WorldSID, female crash test dummy, male crash test dummy, occupant protection, frontal impact, side impact, vehicle ratings, automotive safety

Transcript Highlights:

• This committee will gather relevant local and state information and, as needed, feed those insights to
• public display of a minor's list of friends, contacts, or followers, and prohibiting the use of algorithms
• The algorithms used by these platforms are... growing kids to the extent that I'm having kids staying
• And then banning the algorithm recommendations.
• Now it is just the endless scrolling. algorithms that are so smart at keeping your attention and keeping

Bills: SB1065, SB2117, SB2217, SB2363, SB2533, SB2742, SB2743, SB2881, SB2964, SB3013, SB3014

Keywords: SB 1065, handgun carry, license holder, concealed carry, open carry, government contracts, governmental entity, property lease, state agencies, local government, higher education, university, Texas Government Code, Penal Code 30.06, Penal Code 30.07, Penal Code 46.03, Second Amendment, gun rights, firearms, civil penalty

Transcript Highlights:

• Social media platforms continue to deny that their algorithms and design features are contributing to
• Their proprietary algorithms and design features are fully capable of stronger safety.
• Just as I feel like the algorithms are responsible for serving a young lady, obviously posts on extreme
• Why do some platforms choose to invest heavily on sending ads and developing algorithms to ensure kids
• To be clear, an algorithm in social media is an artificial intelligence system.

Keywords: 988, house, all

Transcript Highlights:

• is not a ban on algorithmic pricing. is not a ban on algorithmic pricing.
• Then, they put that data into an algorithm, and that data in the algorithm are the deciding factor in
• it through this algorithm it through this algorithm and<02:22:25.200> then<02:22:25.440><
• They disclose any algorithm or anything.
• . algorithmically. algorithmically.

Keywords: 981, all

Summary: The House convened, established a quorum, approved the corrected journal, and heard announcements about committee meetings later in the day. The main floor action was a motion to place House Bill 1101, House Bill 1193, Senate Bill 118, and House Bill 1210 on the special orders calendar, which was adopted without objection. The House then moved into committee of the whole for consideration of House Bill 1101, a bill concerning criminal offenses related to critical infrastructure components and commodity metals. During debate on HB 1101, the sponsor explained that the bill is intended to address theft of critical infrastructure such as street lights, RTD equipment, cell towers, and similar materials by targeting the scrap metal market where stolen goods are sold. The committee adopted amendment L004, which reorganized the statutory structure and aligned offenses more closely with the conduct involved; the sponsor said this reduced the highest charge level from an F5 to an F6 in some circumstances. Amendment L005, a technical correction to L004, was adopted. Several members raised concerns that the bill could unfairly burden scrap yards and legitimate businesses, while others supported the measure as a response to widespread infrastructure theft. The committee rejected amendment L006, which would have required buyers to know or reasonably know that material was unlawfully obtained, with the sponsor arguing it would increase liability and undermine negotiated protections for scrapyard dealers. Amendment L008, which sought to add a good-faith protection for legitimate businesses and reduce chilling effects on commerce, was also defeated after members said similar protections already existed in the bill. Amendment L009, creating an affirmative defense for documented commercial transactions and record-keeping compliance, was likewise rejected because members said the bill already contained a stronger affidavit-based safe harbor. Finally, amendment L10, which would have added a five-year repeal/sunset date for the bill, was debated at length but was not adopted; opponents argued criminal provisions should remain stable, while supporters said a sunset would allow review of the bill’s effectiveness and unintended consequences.

Transcript Highlights:

• The board also recently approved and implemented a discipline algorithm and is moving forward on the
• We currently follow an algorithmic process in the form of a sequence of questions to evaluate complaint
• When I hear algorithms, it's kind of scary because they can have racism and sexism built into them.
• They said, "Well, let me see your algorithm." That's proprietary.
• And that's going to feed directly to their emergency department track board.

Transcript Highlights:

• The accuracy and false positive rates depend on the screening algorithms and cutoffs chosen.
• But when confirmatory DNA testing is included in the algorithm, the false positive rate is established
• You could have just said zero, but you had to throw algorithm in there because you're an AI man.
• Babies who are not fed human milk are at a higher risk of developing NEC, meaning if you are feeding
• donate, ensuring an adequate supply of donor human milk to gather these optimized vulnerable infant feedings

Bills: HB321, HB 1027, HB 1295, HB1782, HB2295, HCR29, HCR56

Keywords: Medicaid, child health, express lane option, income verification, supplemental nutrition assistance, telepharmacy, healthcare access, remote dispensing, rural clinics, pharmacy regulations, health literacy, state health plan, health care, patient outcomes, public health, economic impact, healthcare, breast cancer, cervical cancer, insurance eligibility

Transcript Highlights:

• Senate Bill 815 prohibits the use of artificial intelligence algorithms as a basis to deny, delay, or
• These AI pattern-based algorithms routinely auto-deny approving services and paying legitimate claims
• Computer algorithms.
• And imagine an algorithm, and this is not a traditional...
• A computer, if-then basic programming kind of an algorithm.

Bills: SB 6, SB6, SB504, SB765, SB815, SB929

Keywords: hemp, consumable hemp, hemp-derived cannabinoids, CBD, cannabidiol, CBG, cannabigerol, delta-8, delta-9, intoxicating hemp, hemp gummies, hemp vape, edibles, cannabinoid regulation, hemp licensing, retailer registration, product registration, QR code labeling, child-resistant packaging, minor access

Transcript Highlights:

• When you design these algorithms, do you have classes of problems that you simply don't address?
• errors. ...algorithmic errors that exist within the process you're using for modeling.
• Are they only using algorithms that have gone through your sort of ethical testing?
• identify a putative relationship—some opportunity for efficiency—because when I built... ...these algorithms
• Now, some of the AI packages, I understand, can take the feed, almost write it to modern code, analyze

Keywords: artificial intelligence, waste reduction, fraud prevention, government efficiency, improper payments, data reliability, oversight

Summary: The meeting was chaired by Chairman Schweikert and involved a comprehensive discussion on how to utilize artificial intelligence (AI) for reducing waste, fraud, and improper payments within federal programs. Key witnesses, including Mr. Andrew Canarsa from the Council of the Inspectors General, provided insights on the potential of AI in enhancing government efficiency. The committee emphasized the importance of reliable data and thorough examination of AI application to avoid unintended consequences while addressing the estimated $162 billion in improper payments reported by the federal government. Concerns were raised regarding the recent firing of inspectors general and the impacts that could have on oversight and accountability processes.

Transcript Highlights:

• Algorithms. I don't think people are sitting back there looking at comps.
• algorithms being used in every single industry.
• And so participation and, you know, you buy into the algorithm, you're using the algorithm, ...and so
• participation and, you know, you buy into the algorithm, you're using the algorithm, you're participating
• Employers used to have to monitor video feeds to track workers.

Summary: The committee heard several bills and took up multiple votes. AB 576, dealing with updates to California’s charitable fundraising platform registry after AB 488, was presented as a technical fix to help charities and platforms more quickly update eligible listings; there was no opposition testimony, and the measure was later moved on a due pass basis. AB 1160, on law enforcement drone programs, was presented with proposed amendments requiring U.S.-based data storage and limits on sharing or selling collected data; supporters said it would protect sensitive drone-collected information while preserving affordable drone use, while there was no opposition testimony, and it was also later passed out on a due pass as amended basis. AB 75, which would require insurers to give notice before collecting aerial images of homes and provide homeowners access to those images, drew support from the Department of Insurance and consumer advocates, while insurers opposed it unless amended, arguing aerial imaging is a cost-saving inspection tool and that the bill could worsen availability and affordability; the bill was moved to Appropriations on a due pass basis, with the roll held open for absent members. The committee spent the most time on AB 325, which would address algorithmic price-fixing and collusion. The author and supporters argued that competitors using the same pricing software or algorithm to set prices can function as collusion even when the agreement is hidden behind code, and cited examples involving housing, frozen potatoes, gas pricing, and other industries. Opponents, including apartment, business, hospital, retail, chamber, and tech groups, said the bill was too broad, could sweep in lawful software and public-data market research, and might chill legitimate pricing tools, especially for small businesses and housing providers. Committee members raised concerns about clarity, intent standards, and whether the bill should focus more narrowly on nonpublic competitor data; despite those concerns, AB 325 was moved to Appropriations on a due pass basis, with the roll held open. Later, AB 1221 was presented as a workplace surveillance bill that would prohibit certain invasive or discriminatory surveillance tools, require notice to workers, limit sharing of worker data, and require corroborating evidence before discipline based on surveillance outputs. Labor and consumer groups supported the bill, saying modern surveillance technologies can track speech, movement, emotion, and other sensitive traits and can be biased or abusive. The Security Industry Association opposed it, warning that the bill’s broad definitions could restrict legitimate security systems, emergency sharing with first responders, and employer responses to misconduct or unsafe behavior. The transcript cuts off before any final vote on AB 1221 is shown.