AI Agent Safety Benchmarks
Overview
AI agent safety benchmarks test whether systems using tools, software, or physical devices follow instructions, respect permissions, avoid harmful side effects, and know when to abstain. Common approaches use executable sandboxes, simulated workplaces and embodied environments, multi-turn red teaming, and analysis of full action trajectories rather than final answers alone. Evaluations increasingly measure safety alongside task success, usability, authorization, and recovery, while runtime guards and deterministic checks are tested as possible safeguards. Results depend heavily on the model, harness, environment, and scoring method, and fragmented benchmarks make safety rankings difficult to compare; a central challenge is capturing long-horizon and real-world risks without confusing safe behavior with inability to act.
Latest papers
240 papers29 papers in the last four weeks, down 41% on the four weeks before. 0.2% of all new papers.