Trump Is Killing State AI Laws. Here's What Developers Need to Know.
Quick summary
Trump's executive order directs the FTC to preempt state AI laws and conditions $42B in federal funding on states repealing AI regulations. California, New York and 15 other states are affected.
Read next
- 1,100 Ships GPS-Spoofed: Iran Switches to BeiDou, Apps Break
- AI Agents Will Own More Crypto Wallets Than Humans: Coinbase x402 Is Already Live
President Trump signed an executive order in December 2025 directing the FTC to classify state-level AI bias laws as federally preempted, and conditioning $42 billion in broadband infrastructure funding on states repealing AI regulations the administration labels as onerous. The March 11, 2026, deadlines inside the order are now active. If you build AI products used in the United States, the compliance landscape you planned for may no longer exist.
What the Executive Order Actually Does
The order, titled "Ensuring a National Policy Framework for Artificial Intelligence," declares it US policy to achieve "global AI dominance through a minimally burdensome national policy framework for AI." The operative mechanisms are three:
FTC preemption theory. The order directs the FTC Chairman to issue a policy statement explaining when state laws that require AI developers to alter model outputs to reduce bias constitute deceptive trade practices under federal law — and are therefore preempted by the FTC Act. The underlying argument is that forcing an AI to modify truthful outputs is itself deception. Legal commentators have described this theory as untested and unlikely to survive judicial challenge, but it is now official FTC guidance.
Commerce Department blacklist. The Secretary of Commerce must publish a list of state AI laws considered burdensome impediments to innovation. Once listed, those laws become targets for federal legal challenge through the DOJ AI Litigation Task Force.
$42 billion funding condition. The administration is conditioning previously allocated broadband infrastructure funding on states agreeing to repeal laws on the Commerce Department's onerous list. This is the most direct leverage mechanism — states that want federal broadband money will face pressure to eliminate their AI laws.
Which State Laws Are in the Crosshairs
The universe of affected legislation is broad. Here are the laws most likely to end up on the Commerce Department's list:
California:
- SB 53 (Transparency in Frontier AI Act) — requires frontier AI developers to publish safety documentation and conduct third-party audits
- AB 2013 (Generative AI Training Data Transparency Act) — requires disclosure of training data sources for generative AI models
- AB 1008 — classifies certain AI-generated data as personal information under the CCPA
New York:
- RAISE Act — requires impact assessments for high-risk AI systems in employment, housing, and credit decisions
Other states:
- Deepfake disclosure laws active in Texas, Virginia, Georgia and 12 other states
- Automated decision-making employment laws in Illinois and New Jersey
- Chatbot disclosure requirements in California, Colorado and Connecticut
What This Means If You're Building AI Products
For developers and AI product teams, the practical question is: which set of rules do I have to follow?
The honest answer right now is: both, and you will need a lawyer to tell you for how long.
The executive order does not immediately nullify state laws. Federal preemption requires either explicit congressional action, a court ruling, or a regulatory determination that survives legal challenge. The FTC's new guidance is a regulatory determination, but it will be challenged in court by California and other states, and that litigation will take 12 to 24 months to resolve.
In the meantime, companies building products used in California still face California law. The state has not repealed SB 53 and has no obligation to do so until a court orders it. Similarly, New York's RAISE Act remains on the books.
The practical implication for developers: do not assume state AI laws are gone. Build compliance architecture that can accommodate either outcome — federal preemption or state law survival. The worst position is to drop state law compliance now and then face enforcement if preemption fails in court.
The DOJ AI Litigation Task Force
The executive order establishes an AI Litigation Task Force within the Department of Justice, operational from January 10, 2026. Its mandate is to challenge state AI laws on three grounds: unconstitutional burden on interstate commerce, federal preemption by existing regulations, and other unlawful grounds as determined by the Attorney General.
This is not hypothetical. The DOJ task force has already been active for two months. Expect the first lawsuits against California and New York AI laws to be filed within the current quarter.
The Commerce Clause argument — that state AI laws unconstitutionally burden interstate commerce — is the stronger legal theory. Courts have historically been receptive to arguments that state laws creating a patchwork of requirements across 50 states impose undue costs on companies operating nationally. The preemption argument based on FTC authority is weaker and more novel.
What the Order Explicitly Protects
The final text of the executive order includes carve-outs that limit its scope:
- Child safety laws are explicitly preserved — state laws regulating AI use around minors are not targeted
- State government procurement rules for AI remain intact — states can still regulate how their own agencies buy and use AI
- Data centre permitting and infrastructure regulations are preserved except for generally applicable permitting reforms
- Consumer protection laws based on deception and fraud remain in effect
For developers building in the consumer space: child safety compliance (COPPA, state age-verification laws, deepfake laws targeting minors) remains mandatory regardless of the executive order outcome.
Why This Happened Now
The immediate trigger is California's SB 53, which requires frontier AI companies to publish safety testing documentation and allow third-party audits. The major AI labs — OpenAI, Anthropic, Google DeepMind — opposed SB 53 on the grounds that it would expose proprietary safety research methodology to competitors and foreign actors.
The broader context is that 18 states passed some form of AI regulation in 2025, creating the patchwork compliance problem that the AI industry has been lobbying against for two years. The Trump administration's executive order is the federal response to that lobbying — centralising AI governance at the federal level and eliminating the state-by-state compliance burden.
The irony is that the EU AI Act, which the order implicitly competes with, is simultaneously being enforced across Europe. US companies building globally now face a three-way compliance challenge: EU AI Act (in force), US federal AI policy (in formation), and surviving state laws (in litigation).
The Developer Compliance Checklist
Given the uncertainty, here is a practical framework for AI product teams:
Do not drop state compliance yet. SB 53 and the RAISE Act are still in force. Maintain your existing compliance posture until a court orders otherwise.
Document your training data sources now. AB 2013 (California training data transparency) is likely to be challenged but may survive. The documentation burden is not enormous and the legal risk of non-compliance in California is.
Watch the Commerce Department list. When the list of onerous state laws is published, you will know exactly which laws are targeted. That is when to reassess your compliance architecture.
Build modular compliance. Design your AI pipeline so that bias mitigation, audit logging, and transparency documentation can be enabled or disabled by jurisdiction. The ability to switch compliance features on and off by market will be valuable for the next two years regardless of how the litigation resolves.
The EU remains non-negotiable. Nothing in the US executive order affects EU AI Act obligations. If your product has EU users, you comply with the EU AI Act regardless of what happens to California law.
Key Takeaways
- Trump's December 2025 executive order directs the FTC to preempt state AI bias laws and the DOJ to sue states enforcing them
- $42 billion in broadband funding is conditioned on states repealing laws the Commerce Department labels onerous — this is the primary economic lever
- California SB 53, AB 2013, New York RAISE Act and 18 other state laws are in the crosshairs — but remain in force pending court rulings
- Legal timeline: 12 to 24 months before courts resolve preemption — do not drop state compliance now
- Child safety, state procurement, and consumer fraud laws are explicitly preserved — carve-outs limit the order's scope
- For developers: maintain existing compliance posture, build modular compliance architecture, and watch the Commerce Department's onerous-laws list for your specific exposure
Free Weekly Briefing
The AI & Dev Briefing
One honest email a week — what actually matters in AI and software engineering. No noise, no sponsored content. Read by developers across 30+ countries.
No spam. Unsubscribe anytime.
More on ai-policy
All posts →1,100 Ships GPS-Spoofed: Iran Switches to BeiDou, Apps Break
GPS spoofing put 1,100 ships at airports and nuclear plants in 2026. Iran switched to China's BeiDou, abandoning US GPS. What breaks and how developers build resilient location services.
AI Agents Will Own More Crypto Wallets Than Humans: Coinbase x402 Is Already Live
Brian Armstrong says AI agents will soon make more transactions than humans. They cannot open a bank account — but they can own a crypto wallet. Coinbase already launched x402 Agentic Wallets with 50 million transactions processed.
QuitGPT: 2.5 Million People Are Leaving ChatGPT After the OpenAI Pentagon Deal
The QuitGPT boycott launched after OpenAI signed a Pentagon contract on February 28, 2026. Over 2.5 million people pledged to cancel ChatGPT. Claude surpassed ChatGPT in the US App Store for the first time. Here is what actually happened and what it means.
Studios Want to Own Actor Likenesses Forever. SAG-AFTRA Is Fighting to Stop It.
AI can now clone any actor's face, voice, and performance. Studios are negotiating to own those clones as permanent intellectual property. SAG-AFTRA is pushing back with AB 1836, the Tilly Tax, and CAAVault. Here is what the 2026 negotiations actually mean.
Free Tool
Will AI replace your job?
4 questions. Get a personalised developer risk score based on your stack, role, and what you actually build day to day.
Check Your AI Risk Score →Written by
Abhishek Gautam
Full Stack Developer & Software Engineer based in Delhi, India. Building web applications and SaaS products with React, Next.js, Node.js, and TypeScript. 8+ projects deployed across 7+ countries.