The BaseResearchMETR
Review of the "Risks from automated R&D" section in Anthropic's Risk Report (August 2026)
METR reviews the automated R&D section of Anthropic's August 2026 Risk Report. It agrees catastrophic risk from July 15 models was low, but argues the public report's claim that the 2x acceleration threshold was not crossed is overstated.
Why it matters
The same thing, for different readers.
What it means for you
An outside reviewer says Anthropic's public report made AI-driven speedup look smaller than Anthropic's own private report suggested.
For lawmakers
METR says threshold claims in a company's public safety report relied on unclear judgment calls, which supports requiring outside review with access to private versions.
For researchers
METR notes the figure implied by the private report is 1.97, not clearly 'less than 2', and suggests alternative wording ranging from 1.3 to 2.8.
The question it raises
When a company grades its own risk, how plain must it be about the judgment calls that decide whether a warning line was crossed?
In their own words
Every quotation was found, word for word, in the text we saved from the source. That shows the words are there; it does not make them true, or prove the speaker meant them as quoted.
From the source’s own text.
“We agree with Anthropic that catastrophic risks from July 15th models were low, but we think it is plausible that they had crossed their early warning threshold at the time.”
“We believe the public risk report’s conclusion on the 2x Acceleration Threshold seems overconfident, and relies on judgment calls that aren’t communicated clearly by the public report.”
How the finding was reached
In our words, from the source. Weigh the finding by how it was made.
- What was done
- A review of one section of Anthropic's public August 2026 Risk Report, compared with the private version and other non-public data shared during METR's high-access assessment.
- What it found
- The private report's figures imply an acceleration of 1.97, which METR says is better described as 'around 2'; METR's own range for leading indicators is 1.3 to 2.8.
- What it does not show
- Anthropic had rights to redact; some of METR's analysis was redacted from the public version; METR notes methodological uncertainty in measuring acceleration.
- Status
- A report
The record
- Questions
- What happens when AI helps build more capable AI? · Are AI companies putting stronger safeguards into practice? · Who can demand evidence, delay a release or stop an unsafe deployment?
- What we read
- We read the whole source.
- Found by
- The Radar, run of November 6, 2026, via
metr(tier 1: the source’s own channel or an official record) - Cite this
- METR, “Review of the "Risks from automated R&D" section in Anthropic's Risk Report (August 2026)”, November 5, 2026. Entry in the Ctrl AI Base, https://ctrlai.com/base/metr-2026-11-05-review-anthropic-rd-risk-section