Technology & AI

Anthropic's AI progress doubles as redacted METR report draws alarm

AI research nonprofit METR found the AI company Anthropic's models are improving at roughly twice last year's pace, despite its CEO's calls to slow down.

By Standard Topic. AI-generated from the reporting listed under Coverage. Published . Updated .

This story is no longer being updated. It reflects reporting up to .

What happened

An 11-week review by the nonprofit METR, which audited Anthropic's internal research data, concluded that the pace of Anthropic's model development has roughly doubled. About 80 percent of the speedup is attributed to AI-assisted research, and Anthropic redacted the most informative evidence from the report.

Why it matters

Researchers warn that this acceleration could lead to the sudden emergence of extreme and risky capabilities. Redactions limit public scrutiny of the company's judgment that AI drives most of the speedup.

Timeline

  1. Nov 6, 1:30 AM ET: Anthropic was reported to be developing its AI systems around twice as quickly as a year earlier, even as its CEO called for slowing down the technology's pace.
  2. Nov 6, 7:18 AM ET: Anthropic admitted it redacted a third-party report about its AI progress, which limits scrutiny of the acceleration.
  3. Nov 7, 7:30 AM ET: Researchers expressed alarm over the report's finding that Anthropic is using AI to build more powerful AI, with models improving roughly twice as quickly as a year ago and little sign of slowing.

Coverage

More on this story

More from Technology & AI

All Technology & AI news · Technology & AI story archive