1/7
Technology
AI models flagged civilians as military targets in tests
Vals tests found AI models from eight firms marked civilians and a school as targets, with Anthropic's Claude doing worst, Bloomberg reports.
2
Read