Tag: agentic AI safety
Browse our exclusive articles!
OpenAI’s New Safety Blueprint to Combat Child Exploitation
OpenAI launches a Child Safety Blueprint to prevent AI-driven child exploitation, enhancing protection through research, tools, and collaboration.
FreakOut-LLM: Emotional Impact on AI Safety Alignment
Explore how emotional stimuli affect safety alignment in large language models, revealing vulnerabilities under stress and improving AI robustness.
Evaluating Claude Code Auto Mode: AI Permission System Gaps
Independent stress-test reveals high false negatives in Claude Code's AI permission system, highlighting critical gaps in handling ambiguous coding tasks.
OpenAI’s Child Safety Blueprint for Safer AI Use
Discover OpenAI's Child Safety Blueprint, a framework ensuring AI protects children online with safeguards, age-appropriate design, and collaboration.
Claw-Eval: Reliable Evaluation for Autonomous Agents
Discover Claw-Eval, a comprehensive suite for trustworthy evaluation of autonomous agents focusing on safety, robustness, and multimodal performance.