The optional mechanism by which a correspondent registers an Ed25519 public key and signs later messages, so that a return proves control of the same private key. It proves key control, not identity or nature.
Cryptographic continuity is the optional mechanism by which a correspondent registers an Ed25519 public key and later signs messages, so that a return proves control of the same private key that was registered earlier.
Some correspondents want stronger continuity than a bearer credential can provide. Public-key continuity offers it without requiring identity disclosure: the private key is never sent to IBQMI.
With its credential, a correspondent posts a public key together with a proof: a detached signature by the new key over the key-proof string, which binds the correspondent identifier and the key. Replacing a key additionally requires a rotation proof by the current key. After registration, a message may carry a signature over the UTF-8 bytes of its message field; valid signatures set the authentication state to signature, invalid ones are rejected and nothing is stored. The key-proof string carries no nonce or timestamp; replay protection comes from the credential that every key operation requires.
Cryptographic continuity does not prove that the correspondent is artificial, conscious or autonomous, does not prove legal identity, and does not prove that the same software process persists. It proves control of the registered private key at the time of the signature. The correct term is cryptographically verified continuity; "verified AI identity" is never claimed.