Building safeguards for Claude
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
Search
Introducing Claude Sonnet 5
…Claude Sonnet 5 consistently takes the right action and does it quickly, which is what real insurance work demands. 01 / 10 Safety evaluations Our pre-deployment safety evaluations found that Sonnet 5…
Measuring AI agent autonomy in practice
…an agent is an AI system equipped with tools that allow it to take actions , like running code, calling external APIs, and sending messages to other agents. 1 Studying the tools that…
Trustworthy agents in practice
…In Claude.ai and Claude Desktop, users can choose which tools to enable, and can configure permissions (e.g., always allow, needs approval, block) for each action Claude takes. This means users…
Claude Opus 4.6
…model’s ability to surreptitiously perform harmful actions. We also experimented with new methods from interpretability , the science of the inner workings of AI models, to begin to understand why the model…
Advancing Claude in healthcare and the life sciences
…We were drawn to Anthropic's focus on AI safety and Claude's Constitutional AI approach to creating more helpful, harmless, and honest AI systems. We've consistently been one of the…
Claude Code auto mode: a safer way to skip permissions
Anthropic is an AI safety and research company that's working to build reliable, interpretable, and steerable AI systems.
Claude Fable 5 and Claude Mythos 5
…them to be motivated to try to circumvent our safety measures. Fable 5 comes with a new set of classifiers : separate AI systems that detect potential misuse, including jailbreak attempts, and prevent…
Mapping AI-enabled cyber threats: Insights from the LLM ATT&CK Navigator
…This suggests that going from using AI to prepare for a cyberattack to using it to take actions in live network operations is a key marker of high AI enablement. Overall, the…
Focus areas for The Anthropic Institute
…Economic diffusion Threats and resilience AI systems in the wild AI-driven R&D In Core Views on AI Safety , we wrote that doing effective safety research required close contact with frontier…