Human Seal
The open standard for governed AI
The Human Seal is the signed, verifiable proof that a real, authorized human approved an AI's world-affecting action. It is the mark of an open standard, the Human Gate, and anyone can check its signature offline and confirm the key against the one we publish.
The control point where a real human, distinct from the AI, must approve a world-affecting action (a payment, a write, a message, a deletion) before it runs. No autonomous path may bypass it.
The act itself: a specific, authorized person deciding to approve or refuse this exact action, having seen its true context.
The signed, tamper-evident receipt that the assent happened: who approved, the exact action (bound by a content hash), and when. Verifiable by anyone, offline.
One worked example, end to end: an AI agent that proposes supplier payments, gated by whichever named person the finance team's own organogram and job descriptions say can approve one, with nothing about the existing accounting system touched.
The Seal's authority does not rest on our word, but it does rest on our published key. Paste any Human Seal into the verifier and confirm the signature in your own browser, offline. To confirm the key is ours, compare it with the key we publish at did:web:humanseal.world (one web fetch). You then rely on that published key, not on what we say about any receipt.
AI has begun to act: to send money, change records, and answer for organizations. The question every board is asking is simple, who said yes? A growing number of jurisdictions and the first binding AI treaty now require a human to be able to oversee and intervene, with a record kept. One Human Seal is that record.
In enforce mode, every world-affecting action is held for a human approval and logged. A new agent's gate starts in shadow mode, which records and holds nothing until a person switches it to enforce.
Each approval produces a signed receipt, bound to the approver and the exact action, verifiable offline.
A present-human, risk-scaled step-up, with a second-approver check for the highest-risk actions. Specified and in design; no system holds it yet.
The Human Gate standard is free to implement. The Human Seal mark is stewarded by HumanGate International (in formation), incubated by FLOCORE and opening to co-founders. We say plainly where the standard is today, in a list of what is and is not built, and we make trust checkable rather than asking for it.