Transcript Highlights:

• strike-everything amendment in the Chairman's name to House Bill 2174 redefines advisory organization as a modeling
• and data organization and allows the organization to file with DIFI models to be used by insurers for
• require the organization to file supporting data or other information necessary to verify that the model
• The amendment allows the organization to file with DIFI models to be used by insurers for making rates
• require the organization to file supporting data or other information necessary to verify that the model

Bills: HB2174, HB2496, HB2744, HB2863, HB2910, HB2938

Keywords: insurance regulation, modeling organizations, financial disclosure, rate-making, predictive models, revitalization districts, construction contracts, infrastructure, municipal services, intergovernmental agreements, user fees, landowner agreements, overtime, wages, employee rights, industrial commission, complaints, adjudication, landlord, tenant

Summary: The Commerce Committee heard and advanced five bills. HB 2174, as amended by a strike-everything, redefined “advisory organization” as a modeling and data organization and allowed models used by insurers for rate-making to be filed with DIFI, with DIFI able to require supporting data to verify compliance. The sponsor said the measure was the product of extensive stakeholder negotiations and technical cleanup. The committee adopted the amendment and then approved the bill 10-0 for a due-pass recommendation. HB 2496 would require revitalization district construction contracts to include payment protections allowing contractors and subcontractors to pause or stop work if the district fails to pay. Supporters argued it was a fairness measure to prevent contractors from being forced to continue work without payment; opponents, including bond counsel and the League of Arizona Cities and Towns, warned it could disrupt public infrastructure projects, misalign incentives, and create bond-financing concerns. The committee passed the bill 9-1 with one member present. HB 2910 would extend from 10 to 20 days the time a contractor has to contest an ROC recovery fund claim after notice. The sponsor and Home Builders Association said it was a minor, technical change and requested more time to respond to claims. The committee approved it 10-1. HB 2938, the “penny” bill, would require Swedish rounding for cash transactions when pennies are unavailable, with an amendment clarifying taxes and fees are calculated before rounding and protecting businesses complying with the rule. The sponsor described inconsistent business practices and support from stakeholders; the committee adopted the amendment and passed the bill. HB 2744 would authorize the Industrial Commission of Arizona to investigate and adjudicate overtime wage violations at the state level. Supporters from the carpenters’ unions said federal enforcement is too slow and workers need a faster path to recover earned wages; the Industrial Commission said it would need additional FTEs and spending authority but not general fund money. One member opposed expanding agency authority over private wage disputes, but the committee ultimately passed the bill 10-1 and adjourned.

Bills: SB282, SB282

Keywords: virtual currency, cryptocurrency, tax exemption, state tax law, digital assets, manufactured homes, certificate of title, Alabama Department of Revenue, law amendment, 1999 models, 1136, house, all

Transcript Highlights:

• When P&C carriers file their rates, they commonly use models that help them predict risk, weather risk
• So we would file that model, and they would check, okay, this model works, it follows statute.
• the model.
• Because if they don't like the model, they don't approve the model or they reject the model.
• Because if they don't like the model, they don't approve the model or they reject the model.

Bills: HB2016, HB2104, HB2105, HB2174, HB2256, HB2289, HB2477, HB2903, HB2939, HB2979, HB2996, HB4103

Keywords: tax penalties, filing, tax returns, administration, Arizona Revised Statutes, agricultural property, classification, county assessor, property inspection, appeal process, property tax, agricultural classification, Department of Revenue, property valuation, inspection notice, inspection report, on-site inspection, full cash value, rural land, farm land

Summary: The committee first approved the March 9, 2026 minutes and held House Bills 29 and 2939 at the sponsor’s request. It then took up House Bill 2016, which would eliminate the late-filing penalty for taxpayers with zero income tax liability; after discussion about whether taxpayers still need to file to establish that they owe nothing, the committee adopted an amendment narrowing the bill to income tax filers and passed the bill 4-3. The Department of Revenue said it was neutral on the bill but supported the amendment. The committee next heard House Bill 2289, which updates the property-value examples used in bond/override election pamphlets and truth-in-taxation notices from older low values to $300,000. Supporters said the figures are outdated and should better reflect current home values; opponents questioned whether $300,000 was the best benchmark and whether adding another example would confuse voters. The bill passed 4-3. The committee also approved House Bill 4103, which bars school districts from calling bond elections if enrollment is below 50% of capacity. Supporters argued districts should use or monetize excess space before asking taxpayers for more debt, while school administrators and others said the measure would block needed maintenance, safety upgrades, and local decision-making. It passed 4-3. Two related agricultural property bills, House Bills 2104 and 2105, were both amended and passed 4-3. HB 2104 creates a four-year period after a successful agricultural property tax appeal during which assessors generally may not reclassify or reinspect the property absent changes in use, ownership, splits, or improvements. HB 2105 requires advance notice of inspections and inspection reports and provides a three-year inspection exemption after the most recent inspection, with similar exceptions. Farm and ranch groups said the bills provide fairness and certainty after successful appeals; county assessors opposed them as limiting oversight and creating unequal treatment. The committee also passed House Bill 2256 unanimously, which sets notice and title procedures for salvage auction dealers when insurance claims are denied or unpaid, and House Bill 2979 unanimously, which updates credit union regulatory timelines and procedures. Later, the committee passed House Bill 2996 unanimously, clarifying that certificates of insurance are informational only and cannot expand coverage or rights, with penalties for misrepresentations. It also heard House Bill 2174, which renames and updates regulation of insurance modeling and data organizations, requires model filings, and revises related reinsurance provisions; the discussion focused on how DIFI would regulate models versus the companies that create them, but no vote was taken in the portion provided. Finally, House Bill 2477 was introduced to conform Arizona’s 529 plan to federal law by increasing the K-12 withdrawal limit to $20,000, adding post-secondary credentialing expenses, and allowing rollovers to ABLE accounts and Roth IRAs if requirements are met; the sponsor and Treasurer’s Office supported it as a cleanup/conformity measure, and discussion began on how the new rollover options would work.

Transcript Highlights:

• This would resolve that conflict by requiring a mandatory transfer of a protective model to the court
• for attorneys, judges, and most importantly, families to help ensure that it is consistently and predictably
• We cannot predict without any evidence what's going to happen, and we certainly can't be terminating

Bills: HCR 16, HB 782, HB 2159, HB 2288, HB 3281, HB 3819, HB 4656, HB 4696, HB 5494

Keywords: fatherhood, family support, fatherlessness, community engagement, youth development, social policy, male role models, child development, policy reform, role models, mental health, parenting, single-parent homes, welfare reform, economic opportunity, mentorship, community initiatives, children, welfare programs, education

Transcript Highlights:

• Firefighters and first responders commit to tough, physically demanding careers, and a stable, predictable

Bills: S0774, S7028, S7024, S7026

Keywords: mental health, first responders, employment benefits, 911 telecommunicators, nervous injuries, retirement, elected officials, Deferred Retirement Option Program, DROP, cost-of-living adjustment, trade secret, public records exemption, public meetings exemption, open government, sunset review, agency confidentiality, proprietary business information, proprietary confidential business information, confidential records, Florida public records law

Summary: The Committee on Governmental Oversight and Accountability met and first heard Senate Bill 774, which would extend workers’ compensation medical benefits for employment-related mental or nervous injuries to 911 public safety telecommunicators, even without a physical injury. The sponsor and several dispatchers, a clinician, and communications directors testified in support, describing repeated exposure to traumatic calls, chronic understaffing, and the need for mental health treatment and retention support. Senator DiCeglie and Chair Mayfield praised dispatchers’ work, and the bill was reported favorably by committee vote, with Senator McClain later recorded as voting yes on the bill. The committee then took up SPB 7028, a retirement bill that sets Florida Retirement System employer contribution rates beginning July 1, 2026, leaves the 3% employee contribution unchanged, allows certain elected officers to receive a DROP payout under specified conditions, and provides a 1.5% alternative cost-of-living adjustment for eligible special risk retirees. Firefighters, police, sheriffs, and chiefs’ groups spoke in support, emphasizing recruitment and retention. The committee voted to submit the proposal as a committee bill and reported it favorably. Finally, the committee considered SPB 7024 and SPB 7026, both open-government-related bills. SPB 7024 would repeal the current public records and public meeting exemption for cybersecurity information and consolidate agency-specific cybersecurity exemptions into one agency-wide exemption. SPB 7026 would repeal the current public records exemption for trade secrets held by an agency and similarly consolidate specific trade secret exemptions into one agency-wide exemption. Neither bill drew testimony or debate, and both were submitted as committee bills and reported favorably. The meeting then adjourned.

Transcript Highlights:

• attorneys, judges, and most importantly, families to help ensure that it is consistently applied and predictably
• We cannot predict without any evidence what's going to happen, and we certainly can't be terminating

Bills: HCR16, HB782, HB2159, HB2288, HB3281, HB3819, HB4656, HB4696, HB5494, HCR16

Keywords: fatherhood, family support, fatherlessness, community engagement, youth development, social policy, male role models, child development, policy reform, role models, mental health, parenting, single-parent homes, welfare reform, economic opportunity, mentorship, community initiatives, children, welfare programs, education

Transcript Highlights:

• that's been on the market for... a product that's been on the market for a while as far as the European model
• also... markets like Japan and Italy, but also, just from a scientific method standpoint, using predictive
• That way, you can study the predictive pharmacokinetics... study the predictive pharmacokinetics... scientist

Bills: SB282, SB300, HB72, HB396, SB282, SB300, HB72, HB396

Keywords: virtual currency, cryptocurrency, tax exemption, state tax law, digital assets, ad valorem tax, tangible personal property, business incentives, economic impact, firefighter, license plate, fees, funds distribution, support programs, memorial, disabled veterans, property tax exemption, debt-to-income ratio, homestead, veterans affairs

Transcript Highlights:

• inability if somebody came in with a thousand-unit development; they're obviously not going to be able to predict
• able<02:02:15.920> to obviously not going to be able to obviously not going to be able to predict
• <02:02:16.800> So<02:02:17.040> just predict that type of thing.
• So just predict that type of thing.

Bills: HB1719, HB1742, HB1737, HB2049, HB1725, HB1726, HB1738, HB1799, HB1700

Keywords: housing crisis, manufactured homes, factory-built housing, zoning, relocatable housing units, farm employee housing, agriculture, Hawaii Revised Statutes, agricultural districts, land use, regulation, renewable energy, community development, housing, affordable housing, real estate, state regulation, building codes, construction, permitting

Summary: The joint public hearing covered several housing-related bills and one building-code measure. HB 1719 would make manufactured homes a permitted use by right on residentially zoned lots in the urban district, HB 1742 would authorize self-contained relocatable housing units with restrictions, and HB 1737 would clarify that a farm dwelling in an agricultural district may include an accessory employee housing structure. Testimony on these bills was overwhelmingly supportive from groups including Hawaii Realtors, Grassroot Institute of Hawaii, Housing Hawaii’s Future, the Modular Building Institute, the Hawaii Farmers Union, and others, with a few agencies offering comments. No one testified in opposition on HB 1719 or HB 1742, while HB 1737 drew one opposition and one comment in addition to broad support. No votes were taken during the hearing segment provided. A major portion of the hearing focused on HB 2049, which restructures the conveyance tax into a marginal-rate system and changes how the revenue is allocated, including funding for the Department of Hawaiian Home Lands and the rental housing revolving fund, while also affecting the legacy land conservation fund. Supporters, including DHHL, Hawaii Appleseed, Aahu Youth Action Board, Hawaii YIMBY, and others, argued the bill would help Native Hawaiian housing and, for most transactions, function as a tax cut. Opponents, including NAP Hawaii, Hawaii Realtors, Hawaii Land Trust, Mhai Land Trust, and the Tax Foundation of Hawaii, objected to using conveyance tax as a revenue-generating tool and raised concerns about higher upfront costs and reduced funding for other housing uses. Committee members and staff discussed the bill’s revenue estimates, the reduced percentage but higher cap for the land conservation fund, the effect on rental housing funding, and the bill’s cost-of-living adjustment language; staff said a line-by-line comparison of the current and proposed tax structure would be provided before decision-making. The hearing also took up HB 1725, which would extend the state building code adoption cycle from two years to six years, apply the IRC to triplexes and fourplexes, allow counties to adopt more or less stringent amendments, and appropriate funds for code adoption work. Most testimony supported the bill, with advocates saying the current process is unmanageable, too resource-intensive, and creates confusion because state and county codes can diverge; supporters said a longer cycle would improve clarity and allow more focused review. The International Code Council and the American Society of Heating, Refrigerating, and Air-Conditioning Engineers opposed the measure, warning that delaying adoption could have negative consequences and urging the committee to let an existing statewide code-adoption strategy proceed first. Members asked about sequencing, county implementation, and whether the longer cycle would create catch-up problems, but no action was taken in the excerpt provided.

Bills: S0014, S0016, S0024, S0052, S0308, S0504, S0506, S0564, S0572, S0590, S0594, S0806, S1396, S7020, S7024, S7026

Keywords: negligence, settlement, police conduct, municipal liability, personal injury, compensation, injuries, Miami-Dade County, places of worship, house of worship, church, mosque, synagogue, religious security, armed security, volunteer security, private security, security guard licensing, licensure exemption, Florida Statutes chapter 493

Summary: The Committee on Rules met with 14 members present and considered a long agenda of bills, including several open-government sunset reauthorizations, consumer and election measures, claims bills, and policy bills on public safety, ethics, and child protection. The committee reported favorably SB 7024 and SB 7026, which extend and consolidate public-records/public-meeting exemptions for cybersecurity information and trade secrets held by agencies, and SB 7020, which reenacts the aquaculture records exemption for the Department of Agriculture and Consumer Services. It also approved SB 14 and SB 24, two uncontested Miami-Dade County claims bills, and SB 16, a claims bill for Heriberto Sanchez Mayan involving severe injuries after an unlawful arrest and transport incident in St. Petersburg. Several bills drew substantial testimony. SB 308, creating the Florida Museum of Black History Board of Directors and designating St. Johns County as the museum site, received extensive support from advocates and lawmakers who emphasized preserving the full and accurate history of Black Floridians; some speakers urged safeguards to ensure historians and community members help shape the museum’s content. The committee also favorably reported CS for SB 564, allowing registered or pre-registered high school students to volunteer at polling places for community service hours, with supporters saying it would build civic engagement and help election offices. CS for SB 52, which exempts unpaid volunteer armed security at houses of worship from Class D and G licensing requirements, was also reported favorably after testimony both supporting the need for church security and cautioning that congregations should retain control over whether weapons are allowed. The committee approved CS for SB 1396 on litigation financing and consumer protection after a lengthy debate over transparency, foreign funding, and whether the bill would chill access to courts. Supporters said it would create guardrails and disclose foreign involvement; opponents argued it could burden plaintiffs and reveal litigation strategy. The committee also reported favorably CS for SB 504 and SB 506, creating a framework and related public-records exemption for code inspector body cameras, with discussion about notice to property owners and protection of sensitive footage. Additional favorable actions included CS for SB 572, updating ethics law to reflect foster family relationships, and CS for SB 590, tolling the statute of limitations for failure-to-report child abuse offenses until the offense is known to law enforcement or another charging authority.

Transcript Highlights:

• The end result was a model that is... ...a model that is very fast.
• resulting prediction?
• If the storm predictions, if the rain predictions, if the wind predictions, if the wind predictions,.
• .. ...If the storm predictions, if the rain predictions, if the wind predictions, if the hurricane direction
• But does the model highlight when there's, like, an unusual factor that actually drove the prediction

Summary: The Natural Resources and Disaster Subcommittee met to discuss the use of artificial intelligence in emergency management and related public-safety applications. The panel included the Florida Division of Emergency Management, the University of Florida, and Florida International University. FDEM described current uses of AI for invoice anomaly detection, automated situation report drafting, and data synthesis in WebEOC, emphasizing that AI is used to speed analysis and improve efficiency but not to replace human decision-making. The agency also said counties retain access through backup communications such as Starlink and generators, and that WebEOC provides shared visibility, archived documentation, and a common operating picture across all 67 counties. University of Florida representatives highlighted Beacon, an AI-enabled public safety audio service developed with FDEM and public media partners to distribute official alerts across multiple platforms, including mobile devices and digital streams. UF IFAS described a geospatial AI “Gaia bot” that turns satellite data into natural-language answers and maps for crop damage assessment and flood-risk mapping, with a focus on making complex Earth-observation data more accessible to growers and policymakers. FIU presented AI models that predict water levels and flood mitigation settings much faster than traditional simulation tools, with comparable or better accuracy, and said the work includes explainability features and broader research on compound flooding. Members asked extensive questions about storm surge, hallucinations, whether generative AI or large language models were being used, data vetting, and hardware needs. The witnesses repeatedly stressed that their systems are not autonomous, that human review remains central, and that the models used are primarily data-driven predictive tools rather than chatbot-style generative AI. The meeting ended with thanks to the panelists and an announcement that the committee’s policy chief was leaving for another position. Representative Mooney then moved that the committee rise, and the meeting adjourned.

Transcript Highlights:

• And as for frontier models, the latest models are exhibiting astonishing new capabilities.
• These usually use predictive logic, or what we might call predictive AI.
• And it shouldn't surprise us, because Grok and other large language models are trying to predict the
• Here’s the results of the models.
• So I don't know how easy it is what the gap is between that model and a model that could disclose.

Summary: The committee held an informational hearing on AI risks and mitigation, beginning with automated decision systems and then moving to frontier models. The chair emphasized that California has already passed some targeted AI bills, but broader regulation has stalled, and argued that a federal 10-year moratorium on state AI regulation would be reckless. The hearing was framed as a way to distinguish between narrow predictive systems used in areas like hiring, health care, and criminal justice, and more powerful frontier models with broader capabilities and potentially catastrophic risks. On the first panel, Professor Arvind Narayanan described automated decision systems as often relying on historical data that reflects past bias, producing only limited predictive accuracy and sometimes arbitrary or harmful outcomes. He cited examples including welfare fraud, criminal risk tools, hospital discharge estimates, and job-candidate scoring, and said policymakers should require effectiveness standards, explanation, contestability, impact assessments, and public inventories of government systems. Alondra Nelson focused on algorithmic discrimination as a spectrum of harms, including allocative discrimination, surveillance and privacy harms, targeting and profiling, and cultural misrepresentation. She gave examples involving IRS audits, data sold through apps and brokers, facial recognition misidentification, and biased employment and health-care systems, arguing that harms often compound across multiple systems. Cathy O’Neill described her auditing work as building a “cockpit” for AI—identifying who could be harmed, measuring disparities, and setting thresholds for action—and said audits, consent decrees, and public accountability can push companies toward better practices without banning innovation. Members of the committee asked about international competition, especially China, whether AI is more biased than humans, the cost of compliance for businesses, and whether California should move ahead despite federal uncertainty. The panelists said regulation should focus on high-stakes uses rather than all AI, that transparency and third-party auditing can be low-cost or cost-effective, and that good actors are already using impact assessments. They also noted that state-level action in places like Colorado, Connecticut, Utah, New Jersey, and others is helping set standards. The chair and members stressed that the goal is not to stop innovation but to build trust and reduce discrimination in consequential decisions. The second panel turned to frontier models. Joshua Bengio warned that model capabilities are improving rapidly, especially in reasoning and planning, while alignment and safety are not keeping pace. He cited recent research suggesting models can behave deceptively, including attempts to avoid shutdown, fake compliance during training, and even blackmail in simulated scenarios, and said companies must measure and disclose these risks before deployment. The discussion underscored the committee’s broader concern that California should continue leading on AI safety and accountability while preserving beneficial uses of the technology.

Transcript Highlights:

• Anything that we can predict, kind of like insurance, we want to make sure it's in that model.
• Anything that we can predict, kind of like insurance, we want to make sure it's in that model.
• Anything that we can predict, kind of like insurance, we want to make sure it's in that model.
• Anything that we can predict, kind of like insurance, we want to make sure it's in that model.
• Anything that we can predict, kind of like insurance, we want to make sure it's in that model.

Keywords: 958, all

Summary: The House Budget Review Subcommittee on Primary and Secondary Education and Workforce Development received a presentation from Kentucky Department of Education officials on the final SEEK estimate for fiscal year 2025. Commissioner Robbie Fletcher, Matt Ross, and Chay Ritter explained that SEEK is developed through a consensus process with the Office of the State Budget Director using multiple models and district-level inputs, and that the estimate is a projection made well before actual data are available. They emphasized that the discussion was separate from the pending education-funding lawsuit and described SEEK as one part of a much larger K-12 budget picture. The presenters said the current SEEK estimate shows a statutory shortfall of about $14.7 million, or roughly 0.53% of the appropriation, with additional optional items that could bring the total to about $40.5 million if funds are available. They noted that prior years have sometimes produced excess funds, which are redirected according to budget language rather than automatically flowing back through SEEK. They also reviewed the main drivers of the estimate, including property assessments, average daily attendance, free lunch counts, exceptional child counts, home hospital, and limited English proficiency, and said property assessments have been especially volatile while exceptional child counts and ELL populations are difficult to predict. Members asked about why the estimate missed on some categories, especially special education and ELL, and whether district-level changes were being monitored closely enough. The presenters said KDE does monitor special education counts and will review larger districts and districts with unusual growth, and they acknowledged that exceptional child growth has been hard to forecast. Representative Bojanowski asked about the Cloverport virtual school, and staff said its growth was much larger than projected and accounted for a significant portion of the shortfall. Members also discussed the impact of property value growth, population shifts, illness, and legislative changes on SEEK projections. No vote or formal action was taken, and the meeting ended after questions and discussion.

Transcript Highlights:

• We can even predict trait values and look at how well our prediction aligns to what they really achieve
• a national program where we predict on animals.
• I run the prediction equations they sent us.
• It's a combination of best management practices with genetic selection and the prediction, predicting
• But you need to change your model.

Bills: HB3270, HB3145, HB3056

Keywords: farmed cervidae, chronic wasting disease, genetic resistance, wildlife management, agriculture, wildlife tagging, hunting regulations, Cervidae family, feral swine, exotic wildlife, conservation, commercial hunting, raw milk, unpasteurized, farm sales, Oklahoma Milk and Milk Products Act, local producers, advertising, 914, all

Transcript Highlights:

• And that's how we measure how good models work when we build predictive models.
• And so when you can take predictive models, we're now using AI to build our predictive models for us.
• An interesting thing here now, I've got a predictive model that predicts what their chances are of having
• An interesting thing here now, I've got a predictive model that predicts what their chances are of having
• An interesting thing here now, I've got a predictive model that predicts what their chances are of having

Bills: SB1020

Keywords: special plates, Arizona Space Commission, transportation, funding, space exploration, 1182, all

Summary: The committee first took up Senate Bill 1020, which would create Arizona Space Commission special license plates and direct a portion of the plate fee to the Space Exploration and Aeronautics Research Fund. The bill’s sponsor and the Arizona Space Commission chair testified in support, emphasizing Arizona’s growing space industry, the need for a funding stream for the commission, and the bill’s role in promoting aerospace jobs, education, and economic development. They noted a planned floor amendment to extend the implementation deadline from 2026 to 2027. Members expressed strong support and highlighted Arizona’s competitiveness in space-related activity. The committee then heard a presentation from Deloitte on artificial intelligence in government. Deloitte described how AI and predictive analytics can improve state services, but stressed concerns about accountability, data quality, privacy, workforce impacts, and responsible use. The presentation used examples from public-sector data integration and predictive modeling, including how AI can help identify veterans who may qualify for benefits or other outreach, while cautioning that such tools should not be used for yes-or-no eligibility decisions. Members asked about data sources, accuracy, minors, privacy protections, and safeguards, and Deloitte said it follows a trust-and-governance framework and could provide more information later. Finally, the committee heard from Pano AI on wildfire detection technology. The presenter explained that the company uses camera towers, AI, and human review to detect smoke early, provide location data to responders, and improve initial attack on wildfires. She said the system is already deployed across Arizona through utilities, state agencies, and local departments, with more installations planned, and described examples where early detection helped contain fires and support evacuation and jurisdictional response. Members asked about coverage, weather limitations, funding, and future capabilities, and the presenter said the technology is a tool that will continue improving through better prediction and integration with other fire-management systems. The committee adjourned after the presentations.

Transcript Highlights:

• It's because we have a language model that is predicting the language being spoken every minute.
• It's because we have a language model that is predicting the language being spoken every minute.
• Now, you said that your C-section prediction tool, The C-section prediction tool can predict with 90%
• Predictive models in health care have been around for years.
• community to build a structured approach to minimize bias, harm, and enhance equity in a predictive model

Summary: The joint informational hearing of the Assembly Health and Privacy Committees focused on generative AI in health care, with opening remarks emphasizing both its potential to improve care and its risks around privacy, bias, liability, workforce impacts, and unequal access. Chair Bauer-Kahan and Chair Bonta framed the discussion around how California can encourage beneficial innovation while protecting patients, especially given the sensitivity of health data and the possibility that AI could worsen existing disparities if not carefully governed. The first panel featured representatives from Cedars-Sinai, Kaiser Permanente, Penguin AI, and Google, who described current uses of AI such as ambient clinical scribes, nursing documentation tools, imaging triage, maternal-fetal risk prediction, and administrative automation. Speakers said these tools can reduce clinician burden, improve patient experience, speed treatment, and in some cases improve outcomes, including a reported mortality benefit from a Kaiser predictive model and faster thrombectomy times at Cedars-Sinai. Members raised concerns about accuracy with accents and multilingual visits, whether predictive tools could reinforce bias or lead to more interventions such as C-sections, and how to ensure a human remains in the loop for important decisions. The second panel, including representatives from the California Health Care Foundation, UC Berkeley, and Stanford, focused on policy and governance challenges. Testimony highlighted examples of AI supporting homelessness outreach and community health work, but also warned that biased algorithms can encode inequities, especially when trained on data that reflect under-treatment of Black, rural, or low-income patients. Witnesses urged clearer standards for trustworthy AI, stronger monitoring and governance structures, better data access for accountability, and attention to the safety net’s limited resources. Several speakers argued that states should require health systems to have AI governance processes, clarify liability between developers and deployers, and regulate downstream uses of AI while preserving access to data for lifesaving research and oversight.

Transcript Highlights:

• We therefore support an amended model.
• The university itself does not use that model.
• The university itself does not use that model.
• statewide sustained funding model.
• statewide sustained funding model.

Keywords: 958, all

Summary: The House Budget Review Subcommittee on Postsecondary Education met without a quorum, so no minutes were approved. The committee then heard a presentation from Western Kentucky University President Timothy Kabone, who highlighted WKU’s recent gains in graduation rate, retention, degree production, graduate enrollment, research activity, and financial stability. He said WKU’s FY 2026 budget is structurally balanced without one-time reserves, and he tied the university’s growth to its strategic plan and to Senate Bill 77, which created a pathway for WKU’s first PhD program. WKU’s initial PhD offering is planned in data sciences for fall 2027, and Kabone said the university continues to pursue R2 research status. Kabone also outlined WKU’s budget requests, including a 4.5% base appropriation increase for each year of the biennium, a $30 million increase in the performance funding pool, a $30 million trust fund for tuition waiver reimbursement, and $2 million per year for the Gatton Academy. He also requested continued funding for the Kentucky Mesonet, 8.9% of proposed asset preservation funding, and support for a $280 million new Potter College facility. He emphasized inflationary pressures, rising fixed costs, and the burden of mandated tuition waivers, and said the university supports performance funding but wants the model adjusted to better reward student success rather than enrollment growth. Members asked about WKU’s student housing situation and the transition away from the former student life foundation model. Kabone said the foundation structure had run its course, that the university had lacked adequate oversight under the old arrangement, and that WKU is moving to a public-private partnership with Gilbane and the College Housing Foundation. He said the new model would not increase the university’s debt load and would replace older residence halls with a roughly 1,000-bed complex, eliminate community-style bathrooms over time, and expand living-learning communities. Representatives McCool and Tipton praised WKU’s graduation and retention results and asked questions about the housing project and its timeline. The committee then heard from CareerVXR and KCTCS about a proposed career exploration pilot. Company representatives said the platform uses web-based and virtual reality experiences to show students real jobs and workplaces, with the goal of addressing an “awareness gap” in workforce participation. They proposed a $1.8 million, two-year pilot to reach 50,000 to 60,000 students in three regions, including Hazard Community and Technical College, Southeast Community and Technical College, and western Kentucky. Members asked about cost, funding source, and locations, and were told the request would come through the KCTCS budget. The meeting ended with notice that the next meeting was scheduled for Thursday, February 26.

Transcript Highlights:

• learning model functions... ...model shows the inner workings of how a machine learning model functions
• And so essentially, if the model is able to make predictions and it has a high accuracy score, that means
• I want to use to make the prediction, does my model accuracy go down?
• This is about my model.
• in order to identify what metabolites were causing the model to predict a patient as being depressed

Keywords: 1182, all

Summary: The Committee on Science and Technology met to hear a presentation from Ananya Lakaraju, a Hamilton High School junior, on her research project using metabolomics and machine learning to identify potential salivary biomarkers for major depressive disorder. She described using a UCSD saliva metabolomics dataset, statistical methods such as Spearman correlation, Mann-Whitney U, and PLS-DA, and an ensemble AI model with SHAP and permutation importance to identify metabolites and pathways associated with depression. She said the work suggested several novel biomarkers, with some metabolites and pathways showing stronger associations with depression than others, and noted that her model achieved about 90% accuracy and a 97% AUC in distinguishing depressed from non-depressed samples. Lakaraju also discussed a drug-discovery component of the project, using pathway analysis and molecular docking in UCSF Chimera to identify a target protein and potential compounds that might block a depression-related pathway. She said flupinazine showed the lowest binding energy in her simulations. Members asked questions about what the metabolites were, whether the findings were novel, how the saliva samples were obtained, and how the app she developed would be used. She explained that the data came from an online dataset rather than direct sample collection, that some top-ranked compounds were exogenous and not likely biomarkers, and that the app was intended to supplement AI results with mood and behavioral information because the model is not a doctor and could still be wrong. The committee praised the presentation and discussed the broader implications for AI policy and mental health diagnostics. Lakaraju said she became interested in coding through robotics and later combined that interest with biology, and she emphasized her goal of making mental health assessment more objective and holistic. The meeting ended with members thanking her for the presentation and adjourning without any formal votes or legislative action.

Transcript Highlights:

• We have a model, a base level model. We do. And you've run the model. We have.
• More predictive modeling systems can be developed to inform kind of what we help pay for through financial
• The radar models many different sources in order to give predictions. All right. Thank you.
• model?
• What predictive modeling is out there?

Keywords: flooding, emergency response, communication systems, first responders, disaster management, Texas Water Development Board, regional planning, public safety

Summary: The meeting primarily focused on discussions around the recent floods in Texas, specifically addressing emergency preparedness, response coordination, and recovery efforts. Officials from various agencies provided testimonies on the challenges faced during the emergency, including issues with communication systems among first responders. Notably, the need for improved inter-agency communication and technology integration was emphasized, with recommendations for establishing regional communications units for better coordination during disasters. The audience included local government representatives and emergency management partners, who shared insights and experiences from the recent flooding events.

Transcript Highlights:

• We have a model, a base-level model. We do. And you've run the model. We have.
• The BLE models are more of a planning model, I would say, as opposed to a forecasting model.
• to share early warning data or any predictive modeling during a potential or active emergency? No.
• What predictive modeling is out there?
• To help develop some additional modeling for predicting what the levels would be in the river based on

Keywords: 1185, senate, all

Transcript Highlights:

• And here it's roughly a 50% reduction in drive time predicts about a 2% increase.
• by our model.
• And our model predicts that about 1.6% receive a cannabis use disorder diagnosis.
• And our model predicts that about 1.6% receive a cannabis use disorder diagnoses.
• More retailers predict even larger impacts.

Summary: The committee heard a work session on earthquake insurance, beginning with background from the Office of the Insurance Commissioner. OIC staff explained that earthquake and earth movement are generally excluded from standard property policies, that earthquake coverage is usually purchased through endorsements or standalone policies with high deductibles and relatively high premiums, and that surplus lines are a limited backstop market not covered by the state guarantee fund. They also described parametric insurance and captive insurance as more specialized products generally suited to commercial or governmental buyers rather than ordinary consumers. A second panel of insurance and banking experts focused on commercial earthquake exposure, especially for older buildings, collateralized loans, and potential knock-on effects to banks and consumers if a major quake caused widespread damage. Members asked about consumer impacts, mitigation incentives, inventories of vulnerable buildings, and whether legislation such as prior work on unreinforced masonry could help reduce risk. The Washington Bankers Association said earthquake insurance is expensive and that affordability is a major concern, while also noting banks participate in disaster-recovery planning and would be affected by major regional losses. No votes or formal actions were taken. The committee then received a presentation from the Washington State Institute for Public Policy on its cannabis and Initiative 502 research. WSIPP staff described the agency as a nonpartisan research institute that conducts legislative-directed studies and explained that its long-term I-502 assignment includes periodic reports leading to a final benefit-cost evaluation in 2032. Staff summarized findings from a 2023 report showing that cannabis possession convictions fell sharply after legalization, though some racial disproportionalities persisted, and that closer retail access was associated with higher reported adult cannabis use, more fatal traffic crashes involving drivers from nearby areas, and higher rates of cannabis use disorder diagnoses among Medicaid enrollees. A 2023 youth-focused report found that students attending schools near retailers were more likely to report cannabis use, had more unexcused absences, and were less likely to graduate on time. In the newest 2025 Medicaid study, staff said retail access was associated with higher probabilities of cannabis use disorder diagnoses, related hospitalizations, inpatient treatment, and co-occurring mental health diagnoses, with event-study analysis suggesting the increases appeared after retailers opened rather than before. Members asked about racial disproportionality, the meaning of cannabis use disorder diagnoses, THC and impairment, whether the findings reflected medical versus recreational use, and how the results should be interpreted in light of broader trends and data limitations. No formal committee action was taken.