Challenge brief

Prevent irreversible AI autonomous actions

Build safeguard that prevents AI agents from taking unauthorized, irreversible actions in the real world setup.

$500

Challenge prize pool

European AI OfficeUK AI Security InstituteNIST

Submit your work

Paste a link to a case study or a working script, such as a GitHub repository.