Trustworthy agents in practice
… Below, we walk through examples drawn from three: human control, alignment with user expectations, and security. Our other two principles—transparency and privacy—run through each. …
Search
Related topics:
Privacy
Measuring LLMs’ ability to develop exploits
… It was developed as a collaboration between UC Berkeley, the Max Planck Institute for Security and Privacy, UC Santa Barbara, and Arizona State University with contributions from security researchers at Anthropic, OpenAI, and Google , as a follow-on to the CyberGym vulnerability-reproduction benchm… …
How people ask Claude for personal guidance
… To preserve people's privacy, we relied on automated graders Claude Sonnet 4.5 , which may miscategorize conversations see Appendix . …
Results from first Anthropic Public Record
… Only two—privacy and child safety—draw outright majority support for more than a minimal role. National security, meanwhile, has the narrowest partisan gap of any domain, just three points between Democrats and Republicans. …
Measuring AI agent autonomy in practice
… Developing these methods in a privacy-preserving way is an important area for cross-industry research and collaboration. …
Anthropic Economic Index report: Economic primitives
… 1 As with previous reports, all our analysis is based on privacy-preserving analysis. …
Claude Fable 5 and Claude Mythos 5
… Cybersecurity . …
How AI Is Transforming Work at Anthropic
… The Security team often uses Claude Code for code understanding 48.9% , specifically analyzing and understanding the security implications of different parts of the codebase. …