Full Breakdown
States and International Leaders Accelerate AI Regulation as Federal Action Lags
By Drooid · · How we work
Core Event: State and Global Moves to Regulate AI
State governments and foreign ministries are advancing policies to oversee “frontier” AI models while Congress remains stalled after the November elections. Oregon issued an executive order directing its chief information officer to devise third-party safety-review standards within 90 days. In Michigan, attorney-general candidates debated AI consumer-protection rules during a televised debate. Rep. Jodey Arrington warned that heavy regulation could jeopardize U.S. competitiveness, while other experts called for clear safeguards. Canada’s AI minister announced G7 and G20 talks aimed at a “pragmatic” global framework, and California and New York have enacted or are preparing state-level safety-assessment legislation.
Background & Context
Federal leaders have repeatedly declined to impose basic guardrails on AI. President Donald Trump has dismissed the need for oversight, a stance echoed by congressional leaders who fear regulation could hinder innovation. In the absence of federal action, states such as California and New York have passed bills requiring independent safety audits and registries for AI developers. Oregon’s order builds on that trend, seeking to define “frontier” models and evaluate a “kill switch.” Internationally, Canada is coordinating with G7 and G20 partners after recent AI-related data breaches, underscoring cross-border risk.
Key Figures & Groups
- Gov. Tina Kotek – Oregon, author of Executive Order 26-26.
- Rep. Jodey Arrington – Texas, advocating minimal regulation.
- Eli Savit – Michigan AG candidate, proposing AI rules under the Consumer Protection Act.
- Doug Lloyd – Michigan AG candidate, emphasizing youth protection.
- Evan Solomon – Canada’s AI minister, promoting coordinated G7/G20 regulation.
- Sen. Jerry McNerney – California, author of SB 813.
- Gov. Kathy Hochul – New York, overseeing the Responsible AI Safety and Education Act.
Official Statements & Responses
Oregon’s order tasks CIO Terrence Woods with drafting criteria for independent safety reviews and a definition of “frontier” AI, and assessing a potential kill-switch mechanism. California’s SB 813, signed by Gov. Gavin Newsom, creates a statewide registry of AI auditors and requires independent risk evaluations, described as a “strong first step.” New York’s upcoming regulations will require developers to register and disclose safety protocols by early 2027, outlined by Gov. Hochul as part of a broader strategy. In Michigan, Savit pledged AI rules within his first 100 days, targeting deceptive marketing and surveillance pricing; Lloyd argued the AG’s role should focus on protecting children from manipulative AI.
Criticism & Opposition
Rep. Arrington warned that “erring on the side of innovation and winning the race is the right initial posture,” contending that over-regulation could erode U.S. competitiveness. Former Indeed CEO Chris Hyams countered that the debate should focus on *how* to regulate, noting that unclear rules stifle companies and voters favor stronger protections.
Conflicting Reports & Gaps
State proposals differ on which AI systems require third-party review. Oregon’s order does not specify the technologies subject to assessment, and the timeline for finalizing standards beyond the 90-day draft remains undefined. No federal definition of “frontier” AI exists, creating uncertainty for multi-state compliance.
What’s Next
Oregon’s CIO must submit a draft safety-review framework within 90 days, after which the governor will decide on formal regulations. California and New York are set to implement their AI safety statutes later this year and early 2027, respectively. Canada’s G7/G20 working groups will continue discussions on a “red-phone” crisis-communication channel for AI incidents. Congress is expected to revisit AI legislation during the lame-duck session, though no timetable has been set.