Proposing the SAFE Working Group: An Open Community Effort to Improve AI Security

There is no broadly adopted community framework for confidentially sharing AI operational failures , identifying recurring control failures , and ...