H.B. 286 Artificial Intelligence Transparency Amendments
2 additions to document , most recent about 1 month ago
| When | Why |
|---|---|
| Feb-10-26 | HB 286: Government Overreach in AI Development |
| Feb-19-26 | Opinion: Techy legislation abounds in the Utah Legislature |
0 General Document comments
0 Sentence and Paragraph comments
0 Image and Video comments
LONG TITLE General Description: This bill enacts the AI Transparency Act relating to transparency and whistleblower protections for frontier artificial intelligence models.
Highlighted Provisions: This bill: ▸defines terms; ▸requires developers of certain artificial intelligence models to create, implement, and publish public safety and child protection plans; ▸requires developers to publish summaries of risk assessments for certain artificial intelligence models; ▸prohibits developers from making materially false or misleading statements about covered risks; ▸requires developers to report certain safety incidents to the Office of Artificial Intelligence Policy (office); ▸requires the office to provide annual assessments and legislative recommendations regarding regulation of certain artificial intelligence models; ▸establishes civil penalties for violations; ▸provides whistleblower protections for employees who report safety concerns of certain artificial intelligence models; ▸establishes remedies for employees who suffer adverse action for whistleblower activities; ▸creates the AI Transparency Enforcement Restricted Account to fund enforcement activities; and ▸provides a severability clause.
Money Appropriated in this Bill: None Other Special Clauses: None Utah Code Sections Affected: ENACTS: 13-72b-101 , Utah Code Annotated 1953 13-72b-102 , Utah Code Annotated 1953 13-72b-103 , Utah Code Annotated 1953 13-72b-104 , Utah Code Annotated 1953 13-72b-105 , Utah Code Annotated 1953 13-72b-106 , Utah Code Annotated 1953 13-72b-107 , Utah Code Annotated 1953 13-72b-108 , Utah Code Annotated 1953 13-72b-109 , Utah Code Annotated 1953 13-72b-201 , Utah Code Annotated 1953 13-72b-202 , Utah Code Annotated 1953 13-72b-203 , Utah Code Annotated 1953 13-72b-204 , Utah Code Annotated 1953
Be it enacted by the Legislature of the state of Utah:
Added February 10, 2026 at 11:04am
by Sean Urwin
Title: HB 286: Government Overreach in AI Development
Artificial intelligence is moving fast, and Utah lawmakers are right to care about safety and kids. But HB 286 takes Utah in the wrong direction. It leans on broad paperwork mandates, public disclosures, and big penalties for developers. The result is predictable. Companies either slow down useful updates, over filter anything that could be controversial, or decide Utah is not worth the risk. That means fewer helpful tools for families, teachers, and teens, especially in areas like learning support and mental health resources.
You’re asking a really specific—and important—question here, and the honest answer is: there isn’t yet a clear public record of how Senator Todd Weiler feels about this 2026 AI bill (HB 286) in particular. The bill is still in play, and I’m not seeing quotes, votes, or statements from him directly tied to it yet (I couldn’t find a verifiable source link showing his stated position on the 2026 “Artificial Intelligence Transparency Act”).
What we can see is a pattern in how he’s handled some related issues:
- On an earlier, unrelated HB 286 (financial literacy in 2019), Weiler actually served as the floor sponsor in the Senate, which shows he’s comfortable carrying House bills he supports and working on technical policy details ((Utah Legislature, H.B. 286, 2019)).
- He’s also been the Senate floor sponsor on several education/curriculum and content-control bills, like 2024’s HB 29 on “sensitive materials” in schools, a measure backed by conservative groups concerned about books and school content ((Alliance for a Better Utah, H.B. 29 Summary, 2024)).
- In tech/online safety, he’s willing to regulate big platforms: he’s the prime sponsor of SB 142, the “App Store Accountability Act,” which pushes strong age verification and parental consent requirements on app stores and developers to protect minors online ((Utah Bankers Association, Legislative Tracking, 2025)).
- He’s also helped carry mental‑health‑related policy, like sponsoring the Social Work Licensure Compact bill that aims to expand access to mental health providers in Utah ((Social Work Licensure Compact, Utah Enacts the Social Work Compact, 2024)).
So while we can’t say “he supports/oppose HB 286” today without making things up, his record suggests he:
• often supports stronger rules around kids’ safety, content, and platforms;
• is comfortable backing fairly aggressive regulatory frameworks when they fit those goals;
• also occasionally pays attention to technical/implementation details (e.g., recodification and compact bills).
If you want to pin him down on this bill, the most effective move would be to:
- Email or call his office with 2–3 pointed questions: e.g., “Will you support HB 286 as written, and how will you protect both kids’ safety and access to useful AI tools for learning and mental health?” and then track whether he answers in terms of kids’ safety only, or also mentions innovation and over‑regulation.
If you had to write him one short paragraph today—based on what you’ve already drafted about HB 286 “taking Utah in the wrong direction”—what’s the one main point you’d most want him to feel, and do you want to click Continue with AI here to sketch that out together, or head over to Create Document to start a full letter or upload a source to annotate?
That is why Libertas Institute opposes House Bill 286 from Rep. Doug Fiefia.
The bill also creates legal and practical problems that Utah does not need. It pressures developers to publish detailed safety and child protection plans and to keep publishing new summaries as models change. That is not just transparency. It compels public narratives about internal risk judgments, backed by million-dollar penalties. The triggers are also vague. Terms like “severe emotional distress” and “material modification” are not precise enough for a high-stakes enforcement regime. When the rules are unclear and the penalties are severe, the safe move is not better safety. The safe move is over-compliance, slower iteration, and more restrictions for minors. That backfires on the very people the bill claims to help.
Utah should oppose HB 286 because it abandons the state’s best playbook. Utah has been strongest when it focuses on real harms, clear duties, and room to experiment and learn. This bill moves toward a precautionary system that chills innovation first and sorts out the details later. If lawmakers want better protection for kids and the public, they should pursue standards that are outcome based, technology neutral, and workable in the real world. HB 286 is not that bill.
Added February 19, 2026 at 4:00pm
by Sean Urwin
Title: Opinion: Techy legislation abounds in the Utah Legislature
The current legislative session is tackling 21st-century issues. Legislation addressing the impacts of technology, especially artificial intelligence, social media and data use, is a major consideration for lawmakers this year. We touch upon some controversial topics.
HB286 Artificial Intelligence Transparency Amendments would require AI companies to publish reports on risk assessments, provide protections for whistleblowers and, most notably, force them to implement protocols for “child safety risks,” which the bill partially defines as severe emotional distress. This bill had momentum, but President Trump recently asked Utah leaders to abandon it, citing inconsistency with his recently signed executive order directing states not to pass AI legislation that stifles innovation. Will Utah follow the president’s lead or adopt a different approach?
Cowley: Trump’s executive order directs states not to create a patchwork of AI regulations. Utah lawmakers have a love/hate relationship with President Trump, but being called to the principal’s office by The Donald seems to have caught their attention. Our Legislature doesn’t often back away from its crusade against Big Tech, but with numerous other bills in the works, coupled with Trump’s admonishment, HB286 is not mission-critical this session.
Two years ago, Utah created the Office of AI Policy, which houses the AI Lab. The Lab allows businesses to apply for regulatory relief to implement AI technologies safely and in a studied manner. This balances consumer protections and encourages innovation. This is exactly what we need more of — coordination between tech and government, allowing AI to advance responsibly while placing Utah on the map as the hub for this world-altering technology.
Pignanelli: “Regulation should not restrict fundamental concepts like math, which is essential for AI development.” — Ben Horowitz
When regulating communications, the federal government moves fast. Within months of the first telegraph signal in 1844, the U.S. postmaster was given control of this technology. Congress began supervising the internet through the Telecommunications Act of 1996, a year before I sent my first email. These and similar congressional actions were implemented to prevent a patchwork of state laws and foster innovation.
In 2025, President Trump issued the executive order “Ensuring a National Policy Framework for Artificial Intelligence.” The goals are to curb state AI laws, encourage innovation and eliminate “woke” applications. Because this administration mandate is not a statute, it may not survive a legal challenge. Therefore, Utah lawmakers are incentivized to take action that does not completely contradict the president but allows for some framework to exist should the Supreme Court reject the executive order.
SB287 Targeted Advertising Tax places a tax on online ads. This is a high priority for Gov. Cox and many lawmakers to regulate technology companies while funding various adolescent mental health initiatives. However, these fees would immediately be passed on to businesses that use such advertising, which is a hefty number. Could this be one of the major battles in the session?
Cowley: This is a tax on businesses, plain and simple. Proponents of the bill portray it as a tax on big, bad tech companies, but in reality, the financial burden will be passed along to businesses both large and small. Utah’s entrepreneurial spirit has spawned an impressive number of online startups. These companies range from a crafty mom’s Etsy shop to a multimillion-dollar enterprise. Both thrive because of their ability to reach customers through social media and online ads, which this bill would tax, increasing their costs.
Pignanelli: Supporters of the legislation compare this to the taxes on lifestyle vices that promote remedial activities. The most cited example is the surcharge on cigarettes, which funds tobacco cessation programs and cancer research. SB287 uses the tax on targeted advertising to assist child literacy, youth sports, children’s mental health and public park programs. The argument is that social media causes harm that the fees can remediate.
Despite the compelling messaging, online advertising is not tobacco. Almost all businesses and nonprofits that advertise through technology platforms benefit the economy and do not cause harm. Further, the proposed tax is likely preempted by federal law, a critical issue raised by several senators in a committee hearing.
There are numerous other bills dealing with image, data, and other aspects of technology. Where are they headed?
Cowley: HB276 Artificial Intelligence Modifications, sponsored by Rep. Defay, is aimed at addressing the dark and unseemly possibilities of AI. As online sexual abuse becomes more prolific, it’s imperative that we protect victims in the modern age with legal remedies and takedown rights.
Paris Hilton has been advocating for the DEFIANCE Act at the federal level, which allows victims of explicit AI-generated images to sue the deplorable creators. Legislation should punish the true culprits, the creators, not the technological tool being manipulated for perverse purposes. While GROK is the disgusting outlier, intentionally designed to produce filth, most AI tools strive to implement safeguards. The DUI driver is prosecuted, not the distillery or automaker. AI legislation should follow this framework. People, not computers, are the perpetrators.
Pignanelli: I am among the few remaining dinosaurs who review legislation on paper, not screens. Thankfully, Renae and her contemporaries are negotiating good resolutions to these complicated technology matters.
Note: We represent companies and organizations with interests in the issues described above.
General Document Comments 0
- You’ve got a Utah bill called the AI Transparency Act that would start on May 6, 2026 and it’s all about making big AI companies more open and responsible, especially around public safety and kids.
- It only applies to huge AI developers who use a massive amount of computing power and make a lot of money, especially those running chatbots that lots of people (including minors) use.
- These companies would have to: write and post a public safety plan (for big, rare disasters like mass harm or huge property loss), write and post a child protection plan (for harms to minors), regularly test their AI for those risks, use outside evaluators, and publish summaries of what they found and how they’re fixing problems.
- The bill bans lying or leaving out key facts about safety risks, sets up an Office of Artificial Intelligence Policy to collect reports, send safety incident info to government, recommend future laws, and it lets the state fine companies up to $1 million for a first violation and $3 million for later ones, putting that money into a special fund to pay for enforcement.
- Employees get strong whistleblower protections: they can report safety problems (internally or to the state office), must have an anonymous reporting channel, are protected from retaliation and can sue for things like reinstatement and double back pay—unless they knowingly make false or reckless claims.
If you had to write just 2–3 sentences of your own that explain what this bill does and why it matters, what would you say—and do you want to click Continue with AI to draft that now, or go to Create Document to start your own summary or upload a source to work with?
You’re working with two layers here:
• the Utah bill itself (HB 286, AI Transparency Act)
• the bigger debate about AI safety, transparency, and innovation.
I’ll split suggestions into two groups, and I’ll focus on pieces you can actually cite that line up with your summary of the bill and with both sides of the argument.
- To understand the bill and Utah’s approach to AI: The official Utah bill page for HB 286 gives the text, status, and fiscal notes; it’s your primary legal reference for what the Act actually does ((Utah Legislature, HB 286, 2026)).
- Utah context + “middle ground” model: An op‑ed by Gov. Spencer Cox and Commerce Director Margaret Busse explains Utah’s earlier AI Policy Act and their philosophy of balancing innovation with safety; it helps you see how HB 286 fits into Utah’s broader AI strategy ((Cox & Busse, 2025)).
- News coverage of HB 286 (neutral overview): Utah News Dispatch has a detailed article on HB 286 that summarizes the bill, gives the sponsor’s arguments, and includes comments from supporters and industry critics, which helps you see the live policy debate, not just the statute language ((Cabrera, 2026)).
Articles that give arguments FOR passing bills like HB 286
(these are helpful if you want to build a “pass the bill” case: child protection, catastrophic risk, transparency, whistleblowers)
- Supportive policy coverage of a similar law (California SB 53): Vox explains California’s Transparency in Frontier AI Act, which is very close in structure to HB 286: safety plans, catastrophic-risk focus, and incident reporting; the piece outlines the case that transparency and reporting are necessary to prevent worst‑case AI harms ((Piper, 2025)).
- Tech/policy reporting on SB 53: The Verge’s article on California’s AI transparency law walks through why lawmakers and some AI companies (like Anthropic) ultimately supported this kind of regulation—public safety plans, whistleblower protections, and incident reporting without trying to control content or algorithms directly ((Vincent, 2025)).
- Academic / policy argument for mandatory safety reporting: A position paper, “AI Companies Should Report Pre‑ and Post‑Mitigation Safety Evaluations,” argues that frontier AI companies should be required to disclose safety evaluations to regulators, because voluntary disclosures are inconsistent and often too vague for real oversight; this lines up with HB 286’s idea that companies must publish plans and summaries ((Bowen et al., 2025)).
- Evidence that current AI transparency is patchy: “AI Transparency Atlas” reviews documentation for many frontier models and finds big gaps, especially for child safety and catastrophic‑risk information; it gives you empirical backing for the claim that laws like HB 286 are trying to fix real transparency failures ((Mamirov et al., 2025)).
- Supportive local coverage and framing: Utah Policy and Davis County News both cover HB 286 with emphasis on child safety and catastrophic risks, including the Adam Raine/Reid story; they show how supporters frame the bill as narrow, focused on extreme harms, and non‑intrusive on innovation ((Utah Policy, 2026); (Ginos, 2026)).
Articles that help you understand criticisms / arguments AGAINST bills like HB 286
(these support the Libertas-style concerns you’ve summarized: over‑breadth, innovation chill, vagueness)
- Comparative analysis of AI bills and innovation risks: The AI Legislation Tracker from a major law firm flags HB 286 and similar bills, and often discusses industry concerns about compliance costs, innovation chills, and overlapping regimes, which can help you articulate the “over‑regulation” critique more concretely ((BCLP, 2026)).
- Industry concerns about prescriptive state laws: The Utah News Dispatch article already includes TechNet’s argument that HB 286 combines child-safety and frontier-model safety in a way that’s “overly prescriptive” and could discourage AI development in Utah, so it’s a built‑in source for opposition arguments ((Cabrera, 2026)).
- Critiques of similar California law: The Vox piece on SB 53 also presents critics’ worries that the law focuses narrowly on catastrophic risk and may still miss common harms while adding regulatory friction; these points can be adapted as arguments that Utah needs more targeted, outcome‑based rules instead of broad paperwork requirements ((Piper, 2025)).
- Business‑oriented views on Utah’s earlier AI rules: Legal alerts on Utah’s AI Policy Act and follow‑on laws (like Greenberg Traurig’s and Perkins Coie’s client updates) talk about compliance burdens and uncertainty for companies, which you can use as context for why adding another complex regime (HB 286) might push developers away ((Greenberg Traurig, 2024); (Perkins Coie, 2025)).
If you pick one angle to focus on next—child safety, catastrophic risk, or innovation/over‑regulation—which would you most like to read and write about, and do you want to click Continue with AI to talk it through here or go to Create Document to start drafting your own notes and arguments?
on the uploaded document.
0 archived comments