On 2025-11-07 reporting shows the European Commission is weighing proposals to pause enforcement of certain “high-risk” provisions of the EU AI Act—enacted 1 August 2024—and introduce an expected one-year grace period so obligations would begin after August 2027; exemptions and phased-in transparency and registration rules are also under discussion following pressure from U.S. trade officials and large AI vendors (Meta, Google, OpenAI). This shift matters because it delays compliance costs and alters risk-management timelines for companies, affects investment and operational planning, and complicates international regulatory alignment as U.S. states (notably California’s SB 53, signed 2025-09-29) advance safety and transparency rules. Immediate milestones to watch: the Digital Omnibus text (mid–late November 2025) and regulatory guidance on “high-risk” definitions (late 2025).

EU and U.S. moves this fall signal a potential softening of AI rules: the European Commission is reportedly weighing pausing parts of the AI Act—under pressure from U.S. officials and firms such as Meta and Alphabet—including a leaked Digital Omnibus draft that could exempt narrow/procedural uses from high‐risk registration and grant a one‐year grace period for some obligations beginning after August 2027; the AI Act has been in force since August 2024 with key high‐risk duties slated from August 2026. In the U.S., the White House’s July 2025 AI Action Plan urges discouraging state AI laws while a proposed 10‐year House moratorium was removed by the Senate on 2025‐07‐01. These shifts matter for product launches, compliance costs, competitive advantage, and regulatory certainty; the final Omnibus on Nov 19, 2025 and state/federal moves in early 2026 are the next milestones to watch.

In the past two weeks the field of agentic-AI governance crystallized around new technical and policy levers: two research frameworks—AAGATE (NIST AI RMF‐aligned, released late Oct 2025) and AURA (mid‐Oct 2025)—aim to embed threat modeling, measurement, continuous assurance and risk scoring into agentic systems, while regulators have accelerated action: the U.S. FDA convened on therapy chatbots on Nov 5, 2025; Texas passed TRAIGA (HB 149), effective 2026‐01‐01, limiting discrimination claims to intent and creating a test sandbox; and the EU AI Act phases begin Aug 2, 2025 (GPAI), Aug 2, 2026 (high‐risk) and Aug 2, 2027 (products), even as codes and harmonized standards are delayed into late 2025. This matters because firms face compliance uncertainty, shifting liability and operational monitoring demands; near‐term priorities are finalizing EU standards and codes, FDA rulemaking, and operationalizing state sandboxes.

Over late October–mid November 2025, state-versus-federal governance crystallized as the key driver of U.S. AI policy: states passed aggressive measures (Colorado’s AI Act, effective June 30, 2026, mandates prevention of “algorithmic discrimination,” appeals rights and human review; California allows suits against chatbot firms with civil penalties up to $1,000 per violation), and lawmakers introduced or considered over 1,100 AI-related bills nationwide in 2025. A House provision in HR1 sought a 10-year federal moratorium on state AI rules but the Senate voted 99–1 on 2025-07-01 to strike it. At the federal level, the GAIN AI Act (H.R.5885) was introduced on 2025-10-31 to restrict exports of advanced AI chips and require priority access for U.S. persons. These developments accelerate compliance and enforcement tests for firms, may reshape supply chains, and leave Congress’ balance between innovation and protection decisive for next steps.

On Nov. 12, 2025 EU Commissioner Henna Virkkunen said the European Commission will present a digital simplification package on Nov. 19, 2025 proposing AI Act amendments to ease compliance—potentially including a one‐year grace period delaying enforcement of transparency fines until August 2027—after the AI Act entered into force in August 2024 and with high‐risk rules due August 2026; the goal is legal certainty for firms juggling overlapping rules like the DSA/DMA. In the U.S., the FDA’s Digital Health Advisory Committee met Nov. 5–7, 2025 to consider how generative AI therapy tools should be regulated amid state bans/limits (e.g., Illinois, Utah, Nevada) with civil penalties up to $10,000. Separately, ten foundations pledged $500 million over five years via Humanity AI, with grants starting early 2026. Immediate actions to watch: the Nov. 19 EU package and evolving U.S. federal/state rules on AI mental‐health tools.