AI Regulation: EU AI Act, US States, UK and Korea — 2026-09-02
The first week of September 2026 sees the immediate aftermath of the EU AI Act's August 2 enforcement date, with high-risk obligations deferred by the Digital Omnibus while transparency rules bite. In the US, California lawmakers are advancing bills to curb AI in public comment systems, while a federal preemption fight continues to shadow the growing state-law patchwork. South Korea’s AI Basic Act enforcement framework is solidifying, with new guidelines on AI ethics and advertising standards emerging from regulatory bodies.
AI Regulation: EU AI Act, US States, UK and Korea — 2026-09-02
Top developments
EU Digital Omnibus Delays High-Risk Obligations to 2027/2028
Following the general applicability of the EU AI Act on August 2, 2026, the "Digital Omnibus" regulation (EU) 2026/1744 has officially shifted the timeline for high-risk AI systems. Chapter III, Sections 1–3 of the Act are now postponed to December 2027 and August 2028, providing companies additional time to establish governance structures. This legislative pivot was confirmed in German policy briefs noting that while the Act is "generally applicable," the core compliance burdens for high-risk categories have been rescheduled.
California Advances Bills Targeting AI in Public Comment and Jobs
California legislators are moving quickly as the session ends, introducing measures like Senate Bill 1159 to protect public agencies from AI-generated public comments. This bill aims to make it easier for agencies to filter out synthetic civic engagement, addressing concerns about AI distorting democratic processes. Additionally, broader bills tackling AI’s impact on jobs and social media safety remain under debate, signaling a continued state-level push for specific AI guardrails in the absence of federal law.
US State Patchwork Intensifies with Texas and Illinois Laws Live
As of late August 2026, at least five states including Texas (TRAIGA) and Illinois have comprehensive AI laws either in force or effective January 1, 2027. The Center for Democracy and Technology highlights that while no federal law exists, 29 states have passed their own rules, creating a complex compliance landscape. The recent failure of Senator Ted Cruz's moratorium on state AI laws in the Senate budget package has left the door open for further state-level divergence.
Korea’s KFTC Clarifies AI Advertising Verification Rules
The Korea Fair Trade Commission (KFTC) has issued new guidelines clarifying the verification obligations for performance claims in AI product advertisements. These updates specify the grounds for extending submission periods for evidence and shorten the allowable extension time, aiming to reduce consumer deception in the AI sector. This move follows the implementation of the AI Basic Act's subordinate laws earlier this year, marking a shift from broad principles to specific enforcement mechanisms.
Local view
Germany: Policy analysts note that the "slim" supervisory model adopted by the German government, which centralizes oversight between the BSI (Federal Office for Information Security) and the Federal Network Agency, is now facing its first real-world test with the Cyber Resilience Act's 24-hour vulnerability reporting deadline starting in September. The focus remains on innovation-friendly implementation despite EU-wide pressures.
France: French legal experts are highlighting the financial risks of Article 50 of the EU AI Act, which came into force on August 2, 2026. Penalties of up to 3% of global turnover are now applicable for non-compliant generative AI deployments, prompting urgent compliance reviews for companies using models like GPT-5.6 or Claude.
Context & numbers
- EU Compliance Gap: Reports indicate that 78% of companies have not yet started their compliance preparations for the EU AI Act's high-risk obligations, even before the Digital Omnibus delay.
- State Law Volume: 29 US states currently have active AI regulations, creating a fragmented landscape compared to the unified EU approach.
- Code of Practice Adoption: Approximately 190 companies have signed the EU's Code of Practice on Transparency of AI-generated content as of late July 2026.
On the radar
- January 1, 2027: Colorado's replacement AI framework and California’s CPPA ADMT regulations for significant decisions take effect.
- Q4 2026: First CEN/CENELEC harmonised standards for the EU AI Act are expected to be published, though they remain significantly delayed.
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
