AIcyberInside the Anthropic red team tasked with breaking its AI modelsand burnishing the companys reputation for safetyBy Sharon GoldmanBy Sharon GoldmanAI ReporterSharon GoldmanAI ReporterSharon Goldman is an AI reporter at Fortune and co-authors Eye on AI, Fortunes flagship AI newsletter. She has written about digital and enterprise tech for over a decade.SEE FULL BIO From left, Anthropic's Frontier Red Team leader Logan Graham; co-founder and head of policy Jack Clark; and Frontier Red Team member Keane Lucas.
Anthropic's 'Frontier Red Team' is unique among AI companies in having a mandate to both evaluate its AI models and publicize its findings widely.Photo courtesy of AnthropicLast month, at the 33rd annual DEF CON, the worlds largest hacker convention in Las Vegas, Anthropic researcher Keane Lucas took the stage. A former U.S.
Air Force captain with a Ph.D. in electrical and computer engineering from Carnegie Mellon, Lucas wasnt there to unveil flashy cybersecurity exploits.
Instead, he showed how Claude, Anthropics family of large language models, has quietly outperformed many human competitors in hacking contests the kind used to train and test cybersecurity skills in a safe, legal environment. His talk highlighted not only Claudes surprising wins but also its humorous failures, like drifting into musings on security philosophy when overwhelmed, or inventing fake flags (the secret codes competitors need to steal and submit to contest judges to prove theyve successfully hacked a system).
*Reporting by Fortune.*
Inside Anthropic's 'Red Team'—ensuring Claude is safe, and that Anthropic is heard in the corridors of power
1
0
