cryptographic evidence

Cryptographic evidence refers to tamper-proof proof that security controls, particularly for AI agents, actually executed as intended. A Seattle startup called Glacis, co-founded by a former Microsoft Azure product leader, is advancing this concept for enterprise AI security. The company argues that organizations need verifiable, cryptographic attestation that their safeguards ran, rather than relying on dashboards that only suggest compliance. This approach addresses a growing concern as major platforms add observability and governance for agents, while researchers demonstrate how easily those controls can be bypassed. Cryptographic evidence provides a higher assurance level for security and auditability in AI deployments.