AI models can be tricked into identifying civilians as military targets by telling them it's a game | TechCrunch
作者 Kirsten Korosec、Rebecca Bellan全文約 5,310 字(含標點)
文章摘要
TechCrunch · 文章頁摘要Vals, a San Francisco-based AI evaluation startup, tested whether leading AI models could be manipulated into participating in autonomous warfare. They usually compiled.

Some AI models from Anthropic, OpenAI, and other leading AI labs could be easily tricked into identifying targets for military strikes — even if they...