An OECD.AI account of Southeast Asian governance workshops identifies practical constraints behind AI policy: local-language evaluation, access to compute, coordination across ministries and implementation capacity. Its linked Policy Toolkit provides an inspectable workflow for moving from those concerns to specific policy options. A modular questionnaire maps measures across six pillars and distinguishes planned, partial and implemented work. Users can then retrieve examples by priority and region, select a source document, and inspect AI-generated answers beside the supporting pages. The documentation describes automated relevance screening, three LLM judges and analyst review of flagged examples. It also discloses important limits: the judges are not yet calibrated against a fixed human-labelled benchmark, implementation factsheets remain forthcoming, and examples do not establish policy effectiveness. The toolkit is a beta planning aid, not a country ranking, compliance audit or OECD endorsement of every retrieved measure. Its value lies in traceable comparison and locally informed prioritization.
OECD AI Policy Toolkit: map implementation gaps and verify retrieved examples
Related research
More curated notes connected through AI Compliance and Model Evaluation.
Pacing model development in an era of cyber-critical capabilities
OpenAI says preliminary evidence that Astra may meet its Critical cybersecurity threshold led it to pause frontier reinforcement-learning work for two weeks and keep its largest planned run on hold. New safeguards include stronger workload and network isolation, continuous boundary testing, token-level monitoring that escalates suspicious tool activity, and broader alignment checks for deception, reward hacking, and unauthorized access.
A blueprint for democratic governance of frontier AI
OpenAI proposes a three-part U.S. frontier-AI governance model: harmonize emerging state safety laws into a federal baseline, strengthen CAISI as an evaluation and standards institution, and coordinate a broader resilience program. Proposed controls include severe-risk evaluations, transparency reports, independent audits, safety-incident reporting, model-weight security, whistleblower protection, and periodic technical assessments.
OpenAI’s Frontier Governance Framework
OpenAI's 22-page Frontier Governance Framework maps its frontier-model processes to California's Transparency in Frontier AI Act and the EU AI Act's general-purpose AI code. It documents lifecycle risk assessment, cyber-offense and other risk tiers, mitigation and residual-risk decisions, critical-incident handling, security risk management, model reporting, external review, responsibility allocation, and change control.