Emergent/unpredictable capabilities
Named by 2 kept items in this workspace.
Search every workspace for Emergent/unpredictable capabilities. It matches the name as written, so it can find other things that share it.
The items that name it 2
Complex AI models interacting with real-world systems can display behaviours their creators did not foresee, making security risks difficult to fully assess before deployment.
AI agents may develop capabilities not explicitly programmed or anticipated by designers, complicating pre-deployment security assessment
All techniques in this workspace.