Biometric security firm iProov has rolled out a new human approval service to give businesses verifiable oversight of high‑risk actions taken by autonomous AI agents, addressing a key industry demand for stronger safeguards as organisations expand AI use.
Built around the September‑released Human Approval and Presence Specification (HAPS), the service uses iProov’s widely‑trusted liveness‑check technology to create cryptographically secure proof that the correct, genuine person approved a specific AI‑proposed action. It launches with integrations with solo.io’s open‑source agentgateway and authorisation platform Cerbos, its first policy‑engine partner — matching a call from industry players including Microsoft for provable human sign‑off on high‑impact AI steps.
Under the system, when Cerbos rules flag a high‑stakes action, the AI task is paused. A designated approver reviews the request, completes a live‑human verification check, and receives a unique, one‑time digital signature tied exactly to that action. If any key detail changes, approval is voided and the task is blocked.
The open HAPS standard allows other tech providers to build compatible tools. iProov’s CTO Dominic Forrest will demonstrate the solution at AgntCon North America in San Jose this October.
