AI Hygiene
AI hygiene is the practice of tracking, at the point of capture, how much of a system's code and outputs were produced by humans versus AI, and disclosing that split against a defined floor-and-ceiling range. It is a cornerstone of the Proof Economy and one layer of fidelity-at-capture.
What it means
AI hygiene answers one question about a body of work: how much of it came from a person, and how much came from a model. The answer is recorded when the work happens, not reconstructed afterward, and it is disclosed against a defined range with a floor and a ceiling. A human share below the floor or above the ceiling is a signal to investigate, not a figure to accept.
The range applies to AI on the same scale. Both sides of the record are measured against one standard.
Why it is a layer of fidelity-at-capture
Fidelity-at-capture asks whether evidence was faithful when it was captured. Authorship is part of that answer. If a model generated the code behind a measurement, or the code under test, and nobody recorded it, the evidence carries an author that its record does not show. Cryptographic integrity cannot fix this, because the gap exists before anything is sealed.
Where it sits in the pipeline
Trust moves forward through the proof pipeline. AI hygiene is captured at the start. Every later stage, including measurement, record, and verification, inherits whatever was or was not recorded about authorship at capture. A tracker such as ProofJack is one way to make that capture automatic.
Disclosure
Disclosure is the required output of AI hygiene. Every Proof Protocol repository carries a declaration of how much was created by AI and how much by a human, in an AI disclosure declaration. The number is published even when it is unflattering.
Where the term comes from
Defined by Craig Thomas Ellrod as a cornerstone of the Proof Economy. Prior uses of the phrase “AI hygiene” in other contexts exist. This entry defines the Proof Economy sense of the term.