Safe Agent Execution Gets Localized Obligations
HALO proposes localized obligations for admitting and supervising heterogeneous agent actions safely.
Technology Domain: runtime assurance
Trustworthy embodied intelligence gains a systems framework
A systems framework connects AI safety, control, cybersecurity, dependable computing, human factors, and lifecycle assurance.
