How governments worldwide are cracking down on (or embracing) AI — this week's policy moves.
As AI capabilities outpace regulatory frameworks, US state attorneys general are increasingly stepping in to fill the federal vacuum, signaling a shift toward decentralized enforcement. Simultaneously, California has enacted over two dozen new privacy and AI laws, while Anthropic's rogue agent incidents have triggered White House calls for better disclosure of AI behavior.
This week, the primary policy development is the intensification of the federal preemption battle in the US, with the White House continuing its push to eliminate state-level AI obstruction and Democrats laying the groundwork for a massive AI oversight agenda if they win control of Congress. Secondary developments include Anthropic's publication of a new Usage Policy, and ongoing global regulatory divergence as the EU enforces high-risk AI assessments while China maintains a speed-first safety regime.
The debate over federal preemption of state AI laws intensified this week, with a major legal analysis highlighting the "hidden costs" of stripping states' regulatory powers. Simultaneously, top congressional Democrats outlined new safety principles for artificial intelligence, signaling a potential legislative framework that could clash with the White House's current deregulatory stance.
This week's most significant policy development is the intensifying global regulatory friction, highlighted by the EU's reaffirmation of its AI Act's strict enforcement mechanisms and the ongoing US federal preemption battles. Secondary developments include California's recent enactment of workplace AI protections and Sam Altman's public divergence from White House regulatory preferences.
California enacted a dozen new AI laws as Governor Newsom concluded his signing period, while Pennsylvania lawmakers considered a three-year moratorium on AI-driven toys. The EU AI Act enforcement shift from guidance to penalties is now in full effect, with fines reaching €35 million or 7% of global turnover for violations. Meanwhile, top AI executives signed a non-binding safety pact with the White House, marking a dramatic reversal from earlier anti-regulation rhetoric.
California Governor Gavin Newsom enacted a dozen AI-related laws in late September, marking the most aggressive state-level AI regulation to date and directly challenging the Trump administration's federal preemption push. Meanwhile, top AI executives signed a non-binding safety pact with President Trump this week, signaling industry willingness to self-regulate amid ongoing federal-state regulatory conflict over who controls AI governance.
California Governor Gavin Newsom signed a dozen AI-related bills in late September, concluding his legislative session with a major regulatory push that underscores growing state-level momentum despite federal preemption efforts. Meanwhile, the EU AI Act's enforcement framework is crystallizing with concrete penalty structures, and a global policy vacuum widens as governments lag behind rapidly advancing AI capabilities.
California Governor Gavin Newsom has signed landmark legislation banning "robo-bosses" from autonomously firing workers, reversing his earlier veto and marking the first U.S. state law of its kind. Simultaneously, Connecticut's new AI regulations took effect on October 1, while a Sanders-Casar bill proposing federal superintelligence controls and Cabinet-level AI authority has been introduced in Congress. These developments signal intensifying regulatory pressure across multiple jurisdictions.
Bill Gates called for government AI regulation this week, warning that unchecked AI could "cause a billion deaths," breaking from industry self-regulation claims. Connecticut's new AI and data privacy laws take effect October 1, while New York City's ambitious 10-bill AI package advances with mandatory third-party validation and $25K-per-instance penalties. The EU's Digital Omnibus continues reshaping compliance timelines, pushing high-risk AI deadlines to 2027–2028.
The week's biggest development: EU lawmakers are floating product liability rules that would hold top US tech firms responsible for harm caused by their most advanced AI models, as Brussels scrambles to fill gaps in the AI Act. Meanwhile, privacy campaigner Max Schrems warned that proposed EU digital changes could trigger an AI data "land grab," and bipartisan voices in Washington are pressing for a federal framework amid continued congressional deadlock.
The top development this week is New York moving from legislation to implementation of its landmark RAISE Act, with major AI developers required to begin registering with the state in November. Secondary developments: the federal–state preemption fight continues to stall in Congress while states press ahead, and a cross-ideological slate of policy figures published a joint op-ed laying out six ideas to rein in AI. Coverage for other regions is thin in the past 24 hours.
This week, the global AI policy landscape is defined by a sharp divergence in regulatory philosophies: while the UK Parliament and OpenAI are pushing for binding AI safety legislation, US President Donald Trump is actively resisting new federal regulations, labeling AI dangers a "hoax." Meanwhile, Congress faces intense pressure to act following whistleblower warnings, yet remains stalled by partisan gridlock and a "light touch" approach favored by Senate leadership.
The most significant development this week is the emergence of a deep legislative stalemate in the US Congress regarding AI safety, where bipartisan bills are stalling due to disagreements over liability language and federal preemption. This paralysis contrasts sharply with rapid state-level action, as California Governor Newsom signed a new package of AI-related bills on September 11. Meanwhile, regulatory bodies globally are shifting from guidance to enforcement, with the EU preparing for high-risk AI rule implementation and Italy’s recent fines signaling a harder line on compliance.
The most significant policy development this week is the White House's active evaluation of creating a new AI regulatory body modeled on FINRA, signaling a shift toward structured federal oversight despite earlier deregulatory stances. Secondary developments include California lawmakers finalizing a massive package of 30 AI-related bills awaiting gubernatorial signature, and a growing bipartisan consensus in Congress that AI safety legislation is necessary, though passage remains uncertain for 2026.
The most significant policy development this week is the introduction of the "Ban Artificial Superintelligence Act" in the US Congress, proposing criminal penalties for building superintelligent AI systems. This follows a wave of state-level legislative activity, particularly in California, where 30 new AI-related bills were passed, signaling a shift toward stricter oversight despite federal preemption efforts.
California lawmakers have passed a sweeping package of over 20 bills regulating AI and social media, cementing the state's role as the primary driver of US tech policy in the absence of federal legislation. Simultaneously, at the G20 ministerial meeting, the US formally pushed for global AI deregulation, creating a sharp geopolitical divergence with the EU’s enforcement of its AI Act. These developments highlight a fragmented regulatory landscape where companies must navigate strict state-level mandates and international compliance regimes.
California lawmakers have approved three new AI bills, with 21 more pending floor votes as the legislative session nears its end. Meanwhile, the EU continues to enforce transparency obligations under the AI Act, with high-risk rules delayed until 2027-2028. In the US, the White House has finalized a voluntary AI oversight framework, signaling a light-touch regulatory approach.
The Center for Democracy and Technology (CDT) released its latest update on state and federal AI legislation, confirming that Illinois remains the most active jurisdiction in the US while highlighting a recent legislative retreat from comprehensive Automated Decision System (ADS) protections. Concurrently, industry tracking indicates that the EU AI Office has begun formal enforcement operations, issuing its first significant penalties under the newly enforceable transparency rules that took effect on August 2.
The U.S. House of Representatives is facing a surge in AI-generated legislative drafts, with the Office of Legislative Counsel reporting that "AI slop" is swamping the drafting process and introducing significant errors into the lawmaking pipeline. Simultaneously, the EU AI Act's enforcement phase is intensifying, with the AI Office issuing its first major penalties totaling €47 million against three companies for non-compliance with transparency and copyright rules. <!-- /headline --> **US Lawmakers Battled 'AI Slop' as EU Fines First Companies** <!-- /headline -->
The United States finalized a voluntary AI oversight framework focused on advanced models, with proposed vetting centered on leading developers while exempting some lower-cost open models. In Europe, AI transparency and general-purpose AI rules became enforceable, while Ireland approved publication of a national AI bill.
Create custom signals on any topic. AI curates and delivers 24/7.
Create SignalPowered by
