AI Red Team Engineering notes, research references, and field notes on AI systems, responsible AI, governance, and compliance.
Global search

Search AI Red Team

Search across the complete knowledge base.

Enter at least two characters to begin.
↑↓ select Enter open Browse the full research archive
METR · November 5, 2026

METR’s acceleration assessment: distinguish trends from causal attribution

analysis Model EvaluationAI Compliance
Why it matters

METR’s October 2026 Anthropic assessment finds faster progress on measured capabilities, while emphasizing uncertainty about causal attribution and real-world generalization. Much of the evidence informing its conclusions was redacted. Its public methodology separates historical, current and counterfactual growth and tests sensitivity to benchmark modeling choices; the counterfactual is not directly observed.

My takeaway: For an AI productivity or capability assessment, record baselines, alternative explanations and sensitivity analyses. Distinguish measured trends from interview-based attribution and forecasts. Maintain an evidence-access ledger so reviewers can see which conclusions depend on material they cannot inspect.
Keep exploring

More curated notes connected through Model Evaluation and AI Compliance.

OpenAI News · framework

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.

OpenAI News · framework

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 News · framework

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.