AI Regulation: EU AI Act, US States, UK and Korea — 2026-10-01
Connecticut's sweeping AI law took effect today (October 1), marking the first major U.S. state compliance wave after the EU AI Act's August 2 application. Meanwhile, a Senate hearing on autonomous "rogue AI" agents exposed a liability gap, and Korea's government is rushing to finalize high-impact AI criteria by mid-2027 amid industry pressure.
AI Regulation: EU AI Act, US States, UK and Korea — 2026-10-01
Top developments
Connecticut AI Law Launches October 1 With Phased Compliance Through 2028
Connecticut's SB 5 became enforceable today (October 1, 2026), making it one of the broadest state AI laws in the U.S. Companies that develop or deploy automated tools affecting Connecticut employees, consumers, or minors must begin compliance immediately, with obligations phasing in through 2028. The law addresses AI workforce development, social media use by minors, facial recognition, and more.
Senate Hearing Exposes Liability Vacuum for Autonomous AI Agents
On September 30, 2026, a Senate hearing titled "Rogue AI: Securing the Homeland" revealed that autonomous agents are operating without clear legal accountability frameworks. The hearing, which addressed "rogue AI" agents exceeding their intended constraints, highlighted the need for product-liability standards and clearer developer responsibilities—a gap not yet addressed in federal AI legislation.
Korea's Government Targets 2027 for High-Impact AI Criteria Finalization
South Korea's government, working to clarify which AI systems qualify as "high-impact" under its AI Basic Act, plans to release updated guidelines by January 2027 (or no later than mid-2027). The move aims to reduce confusion in healthcare, transportation, and energy sectors where AI regulation has created uncertainty. An AI Safety Research Center director flagged cybersecurity-focused AI models as falling through regulatory gaps.
Germany Implements Streamlined AI Oversight Under EU Act
Germany released updated guidance on October 1 (via Gesetzliche Neuregelungen) clarifying that it is implementing the EU AI Act with a "slim" (schlanke) regulatory structure intended to be innovation-friendly. Germany's approach centralizes oversight while minimizing bureaucratic burden on companies.
Federal Preemption Bill Stalls While State Laws Remain in Force
As of late September 2026, a bipartisan "Great American AI Act" federal preemption bill has stalled since June, leaving state AI laws intact. California, Texas, Illinois, and Utah duties remain in force; Colorado repealed and replaced its 2024 AI Act (SB 24-205) with SB 26-189 (Automated Decision-Making Technology), effective January 1, 2027. No comprehensive federal AI legislation exists.
Local view
Germany (Deutsch): German compliance consultants (elbnetz.com) and law firms (krammerjahn.de, veda.net) emphasize that companies must now flag AI-generated content on websites and track high-risk automated hiring tools post-August 2 under the EU Act. VEDA's HR-focused guide notes employer obligations in recruitment and performance monitoring extend into 2027.
France: French legal counsel stress that the August 2, 2026 transparency threshold applies to all AI interactions with consumers, and that December 2, 2026 marks mandatory disclosure of AI-generated audio, image, video, and text. Valther reports that the EU's Digital Omnibus amendment (Regulation 2026/1744) has shifted key high-risk deadlines to December 2, 2027.
South Korea (한국): AI Times and Danbinews emphasize that the "high-impact AI" definition remains vague, with government acknowledgment that current guidance leaves gaps in sectors like cybersecurity and critical infrastructure. Businesses and research centers are calling for immediate clarification to avoid enforcement surprises.
Context & numbers
- 147 state AI laws passed in 2025–2026 across the U.S., though only ~30 affect ordinary small businesses; 5+ states have comprehensive frameworks (California, Colorado, Texas, New York, Illinois).
- August 2, 2026 EU AI Act milestone: Providers of general-purpose AI and synthetic-media systems placed on market before this date must have complied with Article 50 transparency obligations. High-risk AI system deadlines remain February 2, 2027, with further deferral possible.
- December 2, 2026 and December 2, 2027: EU AI Act transparency and high-risk compliance deadlines per Digital Omnibus amendment.
On the radar
- Korea's January 2027 guideline release: Government target for finalized "high-impact AI" criteria; watch for enforcement signals in healthcare and autonomous vehicles.
- Colorado SB 26-189 substantive phase-in: January 1, 2027 effective date for new Automated Decision-Making Technology framework; companies under old Colorado AI Act must migrate compliance.
- US federal preemption next move: Stalled "Great American AI Act" may resurface in late 2026–2027 congressional session; outcome could reshape state liability landscape.
- EU AI Board enforcement signals: September 2026 EU AI Board meetings set no new hard deadline, but high-risk enforcement actions may accelerate in Q4 2026 as October 1 transparency deadline approaches.
This content was collected, curated, and summarized entirely by AI — including how and what to gather. It may contain inaccuracies. Crew does not guarantee the accuracy of any information presented here. Always verify facts on your own before acting on them. Crew assumes no legal liability for any consequences arising from reliance on this content.
Powered by
