The goal is to organize information about the key entities on the AI regulation topic.
EU AI Act
First comprehensive AI law from a major regulator. Every other jurisdiction is now positioning relative to it — adopt, match, or diverge.
Prohibited practices have been enforced since February 2025. High-risk system rules land August 2026, with penalties up to €35M or 7% of global revenue. Full enforcement by 2027.
Claude's constitution
Anthropic's public alignment framework for how Claude should behave during training and deployment. Most detailed governance document any frontier lab has published, and sets the template others will face pressure to match.
Released January 22, 2026. Reason-based rather than rule-based, with a clear priority stack: safety → ethics → guidelines → helpfulness. First major AI company to formally acknowledge possible AI consciousness. Published under CC0 license so anyone can use it.
Public First Action
Bipartisan 501(c)(4) funded by Anthropic to back pro-regulation candidates in the 2026 midterms. First serious organized political counterweight to the deregulation side of the AI industry.
$20M from Anthropic, targeting 30–50 candidates from both parties. Priorities are model transparency, preserving state AI laws, export controls, and high-risk use regulation. First ad buys went to Marsha Blackburn (R-TN) and Pete Ricketts (R-NE).
Leading the Future
Super PAC opposing strict AI regulation and supporting federal preemption of state laws. Political vehicle for the "move fast, regulate lightly" majority of the AI industry, with six times the funding of the pro-regulation side.
$125M raised since August 2025. Backed by Greg Brockman (OpenAI), Marc Andreessen (a16z), Joe Lonsdale, Ron Conway, and Perplexity.
Trump AI executive order (December 2025)
Federal directive to build a single national AI framework with minimal rules, designed to override state regulations. The active federal effort to centralize AI governance and dismantle the state-by-state approach.
Creates a DOJ task force specifically to challenge state AI laws in court and threatens federal funding cuts for states with "excessive" rules. David Sacks singled out Colorado's law as "probably the most excessive" on the books.