Track · AI in Product4 min read

280 AI Laws in One Year: Why Federal Deregulation Means More Rules, Not Fewer

Washington is deregulating AI, but states passed 280 AI laws this year and the EU is enforcing its Act — CIOs now face more rules, not fewer.

The AI deregulation illusion: Why CIOs face more rules, not fewer
The AI deregulation illusion: Why CIOs face more rules, not fewergeralt / pixabay

Context

  1. 280 AI laws were passed by US states this year, with weeks still remaining in the legislative session.

  2. California's SB 53 took effect in January, requiring frontier AI developers to publish safety frameworks and report critical safety incidents.

  3. Illinois' SB 315, effective January 2027, mandates independent compliance audits for the largest frontier model developers.

  4. The EU's AI Omnibus amendments, adopted July 2026, extended high-risk AI Act deadlines into 2027 and 2028 while keeping the framework intact.

  5. Colorado's SB 26-189 imposes disclosure obligations on automated decision-making in employment, lending and housing from January 2027.

American lawmakers passed 280 AI laws this year alone — with weeks still left on the legislative clock — even as the White House pushes to dismantle regulatory barriers. For CIOs and product leaders shipping AI across state and national borders, federal deregulation hasn't reduced the rulebook. It has fragmented it.

The core problem is jurisdictional arithmetic. An AI application that satisfies requirements in one state may trigger fresh obligations in another, and the same underlying model can power both a low-risk productivity assistant and a consequential hiring tool — identical technology, different rules. As the source analysis puts it: federal deregulation does not equal enterprise deregulation.

What are the states actually doing?

Last December, the White House issued an executive order seeking to limit conflicting state AI regulations and create a consistent national approach. But executive orders don't automatically invalidate state laws, and the states kept legislating.

  • California's Transparency in Frontier Artificial Intelligence Act (SB 53), effective this January, requires covered frontier AI developers to publish safety frameworks and report critical safety incidents.
  • New York's RAISE Act imposes similar obligations starting January 2027.
  • In September alone, Governor Gavin Newsom signed 11 more AI safety bills, capping a legislative year with nearly 30 AI-related measures in California.
  • Illinois' AI Safety Measures Act (SB 315), enacted July 2026 and effective January 2027, adds safety frameworks, serious-incident reporting and — for the largest developers — independent compliance audits.
  • Texas' TRAIGA, in effect this year, stands apart with an intent-based liability model instead of broad pre-deployment risk assessments.

Two more Colorado laws reach directly into operations. SB 26-189, enacted May 2026, replaces the state's earlier AI law with a framework governing automated decision-making in consequential areas — employment, lending, housing. From January 2027, covered businesses face disclosure obligations and consumers gain rights over inaccurate personal data. HB 26-1263 adds disclosure and minor-protection requirements for conversational AI services.

The lesson here: you can't certify a model once and assume every downstream application is compliant everywhere.

What about the EU?

Global complexity compounds the state patchwork. The EU AI Act is moving into active enforcement, and in July 2026 the EU adopted the AI Omnibus amendments, simplifying some requirements and pushing high-risk implementation deadlines into 2027 and 2028. The core framework survived intact: transparency rules, prohibited practices, general-purpose model obligations and high-risk requirements.

Picture a multinational screening job applicants or automating service interactions. The same application could face different requirements depending on where it runs, who it affects and how its outputs steer decisions.

Why can't enterprises see the AI they govern?

Fragmented regulation is exposing a visibility gap. Enterprises may sanction a handful of AI platforms while departments independently wire models into customer service, analytics, software development and decision workflows. Third-party vendors keep embedding AI into existing products, expanding the footprint without matching governance.

Agentic AI raises the stakes: systems that access sensitive data, interact with enterprise applications and take actions rather than just generate responses. When something breaks, teams need to reconstruct which system was involved, what data it touched, how it behaved and which safeguards existed. Shadow AI compounds this — the more tools and workarounds spread, the harder governance gets.

What does a workable compliance foundation look like?

Building a separate compliance process per jurisdiction is unsustainable at 280 laws a year. The recommended approach is a common enterprise-wide foundation for responsible AI, with added safeguards where specific regulations or use cases demand them. That means:

  • An accurate inventory of AI applications, models, agents and third-party integrations.
  • Continuous evaluation, not just pre-deployment testing — changes to prompts, data, integrations or underlying models can introduce new risks after launch.
  • Ongoing monitoring for accuracy, reliability, security, privacy, bias and policy adherence, with documented evidence of testing and remediation.
  • Governance that extends beyond the model: a reliable model still causes damage when connected to sensitive databases or automated workflows without proper access controls.

Crucially, principles like fairness and accountability must translate into measurable technical controls and operational processes — not a folder of policy documents.

Are simpler rules coming?

Federal lawmakers may eventually pass a national framework that supersedes some state requirements, and courts may reshape the landscape. But CIOs can't bank on streamlined rules arriving; even a unified US approach leaves multinationals facing distinct international obligations. The organizations best positioned are those that know where their AI operates, evaluate its behavior continuously and can demonstrate oversight on demand — the conditions AI adoption needs to scale responsibly.

via cio.com (Source)

More from Daniel Okafor

Daniel Okafor

Show full bio

Staff writer covering technology and product news in product management at Roadmap File.

60 articles