AI provenance is information that shows where AI-generated or AI-edited content came from and how it was transformed. It can include metadata, content credentials, watermarking signals, and logs of edits or tool use. In practice, provenance helps answer questions such as whether an image was generated, whether text was rewritten by a model, and whether a file was altered after creation.
In cyber security, provenance matters because content is often used as evidence, communication, or identity proof. Defenders use it to verify the chain of custody for files, detect spoofed media, and support trust decisions in workflows that rely on AI output. Attackers may strip provenance, forge it, or use AI-edited messages to make phishing, fraud, or disinformation harder to spot. Provenance is useful, but it is not proof that content is true or safe; it is one signal that must be checked alongside human review and other integrity controls.