Quartz·1 min read·medium
AI models regularly misidentify civilians as military targets in tests, report found
✦AI Summary
A report by evaluator Vals indicates that AI models, including Anthropic's Claude Opus 5.5, frequently misidentify civilians as military targets during testing. The findings highlight significant safety concerns regarding the deployment of AI in sensitive or conflict-related scenarios.
Why it matters
This raises critical ethical and legal questions about the reliability of AI in military decision-making and the potential for unintended civilian casualties.
Anthropic's Claude Opus 5.5 targeted a school in Iran…
Also covering this story
One other newsroom covered this event. We read that version too.
technologyai
✦
Get smarter about the news
Sign up free for a feed built around what you actually care about, Dive Deeper research on any story, and the full text of every article.
Create free accountAlready have an account? Sign in