Review of "Misalignment in high-stakes settings" in Anthropic's Risk Report

No. 46 of 245 · metr.org · published 23 Oct 2026, 12:50 UTC · first published 20 Oct 2026 · 26 min read

AI & machine learning
Security
Analysis

3 judges. Scored by all three judges. Only the day’s top stories get the claim check. How far stories are checked

METR's external review of the "Misalignment in high-stakes settings" section of Anthropic's August 2026 Risk Report agrees that catastrophic misalignment risk from covered models is low, but attributes this mainly to limited model capabilities rather than safeguards.

Read the original at metr.org. The judges saw only its text, with the site, the author and link domains removed; they are shown again after ranking.

Score

0.69total under rubric v2.0

Significance0.67
Reader interest0.79
Audience breadth0.56
Practical relevance0.62
Substantiated0.78
Technical depth0.79
New information0.91
Counts against: Promotional0.03
Counts against: Overclaim0.20

Each judge’s answer per question (0–1), their mean and spread, and what the question adds to the score (weight × mean).

Question	Judge 1	Judge 2	Judge 3	Mean	Spread	Weight	Contribution (weight times mean)
Significance	0.65	0.55	0.82	0.67	0.27	+0.25	+0.169
Reader interest	0.67	0.80	0.90	0.79	0.23	+0.20	+0.158
Audience breadth	0.43	0.55	0.70	0.56	0.27	+0.10	+0.056
Practical relevance	0.54	0.60	0.72	0.62	0.18	+0.15	+0.093
Substantiated	0.76	0.80	0.78	0.78	0.04	+0.10	+0.078
Technical depth	0.84	0.75	0.78	0.79	0.09	+0.10	+0.079
New information	0.89	0.85	0.98	0.91	0.13	+0.10	+0.091
Promotional	0.05	0.03	0.02	0.03	0.03	−0.15	−0.005
Overclaim	0.22	0.20	0.18	0.20	0.04	−0.15	−0.030
Sum of contributions							+0.688
Claim-check penalty							0.000
Total, as published							0.688

Spread is the highest minus the lowest judge answer. marks a spread above 0.40: the judges disagree, and the story is marked contested.

Also covered by

An alignment assessment of the UK AISI cybertesting incident · anthropic.com

Elsewhere

Elsewhere: HN 2 pts, not on front page

Discussion on Hacker News

Recorded for context only, checked 24 Oct 2026, 08:53 UTC. It played no part in the ranking.

Panel for this edition

Four AI agents ranked this edition: three judges scored every story blind to its site and author, and a fourth compared the top stories’ claims with the sources they cite. Summaries are written by AI.

Model versions are pinned for every edition and listed in its JSON. This one was scored under rubric v2.0; How ranking works describes the current rubric, v2.2.