The Regulatory Tide Goes Out: What Global AI Governance Retrenchment Means for Organizations

In this blog post, authors Jason M. Loring & Michelle Ramsden (Jones Walker) discuss the recent shift from rapid expansion to partial retrenchment in global AI regulation. After an initial wave led by frameworks such as the EU AI Act and Colorado’s comprehensive AI law, developments in 2025 and early 2026 show a consistent pattern of delay, rollback, or reconsideration across multiple jurisdictions, including the United States, Canada, and the United Kingdom.

The authors highlight Colorado as a central case study, where lawmakers are proposing to replace the original AI governance regime with a narrower framework focused on automated decision-making technology (ADMT). The revised approach removes core elements such as mandatory risk management programs and impact assessments, instead emphasizing disclosure, notice, and human review rights, while relying on existing legal regimes to address issues like discrimination. At the same time, the new definition of covered systems may extend to a broader range of tools that materially influence consequential decisions.

In the European Union, the European Commission is not abandoning the AI Act but delaying key provisions and reducing compliance burdens through the proposed Digital Omnibus package. These changes include postponing requirements for high-risk AI systems and narrowing documentation and registration obligations, reflecting both implementation challenges and political pressure to maintain competitiveness. Similar patterns appear elsewhere, including the collapse of Canada’s proposed legislation, the UK’s decision not to pursue comprehensive AI laws, and shifts in U.S. federal policy toward a more decentralized, innovation-focused approach.

Alongside these regulatory changes, the authors describe how enforcement and liability continue to develop through other channels. Sector-specific laws, regulatory agencies such as the Federal Trade Commission, and ongoing litigation are shaping AI governance in practice, even as comprehensive frameworks evolve. In this context, organizations are increasingly turning to standards such as the NIST AI Risk Management Framework and ISO 42001 to structure governance programs, particularly as statutory requirements remain in flux across jurisdictions.

CONTINUE READING…