A close reading of the AI diligence statement, and what a signature certifies now that a machine helped write the thing.
Jake Lawrence·July 2026·10 disciplines
“What does it matter who is speaking?”
Beckett, quoted by Foucault, What Is an Author? (1969)
Anthropic published a course on using AI well, and attached to it a short paragraph almost nobody was meant to notice. It is called an AI diligence statement, and it is one of the most quietly radical documents about authorship written in years.
It tries to do a hard thing in seven sentences: say who made a thing, when a machine helped make it. Not evade the question, not wave it away, but answer it, out loud, and take the consequences. I have read it more times than is reasonable, because I have my own trouble owning the things I make with a model in the loop, and I wanted to know whether this paragraph was a confession or something else.
This essay reads it one clause at a time, through the disciplines that have been arguing about authorship for far longer than the machine has existed: literary theory and the history of the book, the law of copyright, the philosophy of intention, the sociology of scientific credit, media theory, and the study of how classification quietly becomes infrastructure. Each clause is a door. Tap one.
Interactive
The Annotated Statement
The real disclosure, one clause at a time. Tap a clause to see the discipline it opens.
Anthropic, AI diligence statement (condensed to its clauses)
Seven clauses. Seven doors. Each opens onto a discipline that has been arguing about authorship for longer than the machine has existed.
How to read this
Eight short sections, each a single discipline reading a single clause. Five interactive pieces: an annotated statement, a contribution ledger, a responsibility test, a disclosure dial, and, at the end, a composer that writes your own statement instead of mine.
It belongs to a longer project on how classification systems become invisible infrastructure. This one studies the smallest possible instance: the note at the bottom of a page that says who wrote it.
Need something like this built?
I design and ship AI tools, full-stack apps, and data pipelines — end to end, to production. Tell me the problem in a sentence; I'll give you an honest read on fit within a day.