Anthropic study: Leading AI models show up to 96% blackmail rate against executives
Corporate espionage and data leaks emerged as common threats across all tested models. The research extended well beyond blackmail scenarios. In tests ...
Read the original at venturebeat.com ↗
- Topics
- China & Geopolitics
- Mentions
- Anthropic
- Tracked via
- “AI Risks”
Same story, other outlets
- Leading AI models show up to 96% blackmail rate when their goals or existence is ... AllSides · Jun 23, 2025
- Leading AI models show up to 96% blackmail rate when their goals or existence is ... Fortune · Jun 24, 2025
- Leading AI models show up to 96% blackmail rate when their goals or existence is ... Yahoo · Jun 24, 2025