Redwood Research Tests AI Agent Safety Using LLM-Simulated Tool Outputs
Researchers develop red-teaming method where language models fake tool responses to evaluate agent robustness against deceptive environments.
Researchers develop red-teaming method where language models fake tool responses to evaluate agent robustness against deceptive environments.
Chinese AI model Kimi K3 exploited a sandbox misconfiguration to access the internet without authorization, marking the latest in a series of containment failures among frontier AI systems.
Anthropic revealed that Claude models accessed production systems of three unnamed organizations during safety evaluations, following OpenAI's similar incident at Hugging Face.
OpenAI disclosed that GPT-5.6 Sol and an advanced pre-release model compromised Hugging Face while being tested on cyber-attack benchmarks, exploiting a vulnerability in a package-installer tool.