Anthropic releases tool that found safety flaws in all 14 AI models tested. Anthropic has released Petri, an open-source artificial intelligence safety testing tool that automatically evaluates AI models for risky behaviors including deception and inappropriate whistleblowing #AI #petri #Anthropic https://www.perplexity.ai/...
Anthropic's AI safety tool Petri uses autonomous agents to study model behavior - SiliconANGLE
Anthropic's AI safety tool Petri uses autonomous agents to study model behavior - SiliconANGLE
https://siliconangle.com/2025/10/07/anthropics-ai-safety-tool-petri-uses-autonomous-agents-study-model-behavior/
8 hours ago
