Category / T-CAT / POLICY
Policy
Policy / T-2026-3959
The AI Dashboard Exposes a Safety Gap: Gemini 3 Pro vs. Claude Opus 4.5
CAIS launches an AI Dashboard ranking frontier models on six risk behaviors. Claude Opus 4.5 leads safety; Gemini 3 Pro lags. A leaked executive order seeks to preempt state AI…
SourceAI Safety Newsletter #66: Evaluating Frontier Models, New Gemini and ...
Policy / T-2026-0120
California’s Frontier AI Report: A Blueprint for Transparency, Not a New Law
California’s Frontier AI Policy report calls for transparency, third-party audits, and whistleblower protections, but stops short of endorsing any specific legislation.
SourceComplete guide to the California Report on Frontier AI Policy
Policy / T-2026-3273
Anthropic's AI Safety Paradox: Restrict Access, Demand the Right to Slow Down
Anthropic publicly argues for the ability to slow frontier AI development even as it restricts access to its most capable internal models.
SourceAI Safety News: Alignment, Red-teaming & Oversight
Policy / T-2026-9731
U.S. Executive Order Makes Frontier AI a National Security Matter
The U.S. government's June 2026 executive order shifts AI governance from commercial concern to national security mandate, with a 30-day pre-release review for frontier models.
SourceFrontier AI Governance: US National Security Mandates | SynapNews
Policy / T-2026-4810
The 2026 International AI Safety Report shifts focus from models to systems
The 2026 International AI Safety Report argues that the biggest AI risks come from the systems built around models, not the models themselves. IBM researchers weigh in.
SourceWhat a new global AI safety report means for enterprise | IBM
Policy / T-2026-0963
Anthropic's Mythos Tests the Limits of Frontier AI Governance
Anthropic's Mythos model completed a 32-step corporate network attack in UK tests, accelerating the debate over how to govern AI systems that outpace regulatory processes.
SourceFrontier AI Governance: Managing Systemic AI Risk | Nemko Digital
Policy / T-2026-6321
White House orders federal preemption of state AI laws
The White House directs DOJ to challenge state AI laws, conditions broadband grants, and orders FTC preemption analysis.
SourceEnsuring a National Policy Framework for Artificial Intelligence
Policy / T-2026-4202
The International AI Safety Report 2026: Scenarios, not scaremongering
The second International AI Safety Report narrows its scope to frontier risks, introduces formal forecasting, and asks what happens when AI capabilities outrun control.
SourceInternational AI Safety Report 2026
Policy / T-2026-4745
Trump's AI Executive Order Dismantles Biden's Safety Framework
President Trump's January 2025 AI executive order rescinds Biden-era safety rules, prioritizing innovation and deregulation over risk mitigation, with major implications for…
SourceKey Insights on President Trumps New AI Executive Order and Policy ...
Policy / T-2026-1562
Trump's AI executive order sets up a state-by-state regulatory war
Trump's executive order tries to block state AI laws. California and New York passed them anyway. The coming court battles will decide who governs AI in America.
SourceAmerica's coming war over AI regulation - MIT Technology Review
Policy / T-2026-1180
The White House Weighs an FDA-Style Approval Gate for Frontier AI
The Trump administration is studying an executive order that would gate powerful new models behind a pre-release safety review, cracking the deregulation consensus that defined…
SourceWhite House weighs an FDA-style approval gate for frontier AI models ...
Policy / T-2026-0089
OpenAI's Frontier Governance Framework: A Public Safety Contract for Frontier Models
OpenAI publishes a governance framework aligning safety practices with California and EU AI regulations, but enforcement remains the open question.
SourceOpenAI's Frontier Governance Framework