前沿 AI 风险报告（2026 年 2–3 月）
2026 年 2 月，METR 开展了一项试点研究，对前沿 AI 公司在实验室内部使用 AI 智能体时可能出现的不对齐风险进行了评估。Anthropic、Google、Meta 和 OpenAI 参与了这项试点。报告正文分为三部分：
METR395 min read
METR
Stay updated with the latest developments in Artificial Intelligence
Frontier Risk Report (February to March 2026)
A pilot assessment of rogue deployment risk at frontier AI companies. Starting in February 2026, METR conducted a pilot exercise to assess misalignment risks from AI agents used inside frontier AI developers, with participation from Anthropic, Google, Meta, and OpenAI.
METR396 min read
Measuring the Self-Reported Impact of Early-2026 AI on Technical Worker Productivity
A survey of 349 technical workers finds a median 1.4–2x self-reported change in value of work due to AI tools, expected to grow over time, though there are reasons to be skeptical of the magnitude.
METR19 min readReview of the "Risks from automated R&D" section in the Anthropic Risk Report (February 2026)
External review from METR of the "Risks from automated R&D" section in Anthropic's February 2026 Risk Report
METR3 min read
Task Substitution and Uplift
We distinguish three measures of AI uplift -- on old tasks, on new tasks, and in value -- and show that task substitution can cause these to diverge substantially.
METR9 min read
Evidence on AI R&D Progress from NanoGPT
Classifying human and agent contributions to the NanoGPT speedrun, and what publicly tracked challenges can tell us about AI R&D acceleration.
METR8 min read
MirrorCode: Evidence that AI can already do some weeks-long coding tasks
Early results from MirrorCode benchmark with METR: AI agents can complete weeks-long coding tasks, including reimplementing a 16,000-line codebase.
METR1 min read
Fine-tuning experiments on CoT controllability
We find that a small amount of fine-tuning on instruction following in the CoT generalizes to meaningful increases in CoT controllability on an out-of-distribution set of tasks.
METR11 min readRed-Teaming Anthropic's Internal Agent Monitoring Systems
A METR staff member spent three weeks red-teaming a subset of Anthropic's internal agent monitoring and security systems, discovering several novel vulnerabilities.
METR2 min read
Impact of modeling assumptions on time horizon results
Alexander Barry examines how different modeling choices affect METR's time horizon estimates.
METR22 min read
We spent 2 hours working in the future
Thomas Kwa describes a tabletop exercise where METR researchers simulated having access to ~200-hour time horizon AIs.
METR10 min readReview of the Anthropic Sabotage Risk Report: Claude Opus 4.6
External review from METR of Anthropic's Sabotage Risk Report for Claude Opus 4.6
METR2 min read
Many SWE-bench-Passing PRs Would Not Be Merged into Main
We find that roughly half of test-passing SWE-bench Verified PRs written by recent AI agents would not be merged into main by repo maintainers. A naive interpretation of benchmark scores may lead one to overestimate how useful agents are without more elicitation or human feedback.
METR14 min read
Observations from two CLI game reimplementation runs with Opus 4.6
Nikola Jurkovic describes observations from tasking Opus 4.6 with reimplementing Slay the Spire and Balatro in the CLI.
METR9 min read
Showing 32 of 32 stories
You're all caught up.
AiByts Weekly Brief
Get the top AI news and tools straight to your inbox every week.
Weekly · AI news from METR
Join 15,000+ AI enthusiasts