The acquisition rubric for one of the larger contemporary museums in New York leaked last month. A two-page document, internal, dated this calendar year, used by the curatorial committee to score works under consideration for the permanent collection. The document had been redacted on the way out — names removed, attribution scrubbed — but the column headers were intact, and the headers are the story.

There are seven scoring criteria. Five of them are familiar to anyone who has read the museum’s public statements over the last twenty years: art-historical significance, formal innovation, condition, fit-with-collection, scholarly interest. Two are new. The sixth column reads: human authorship. The seventh: machine contribution.

Both columns are scored on a five-point scale. Both are mandatory. Neither is summed; they sit alongside the others, weighted by the curators in deliberation rather than by formula.

What is striking is not that these columns exist — collections committees have been discussing how to handle generative work for at least three years — but that they are scored separately. The committee is not asking, on a single axis, how much of this is a person and how much is a machine. It is asking, on two axes, how much person and how much machine, with the implicit acknowledgement that those numbers can both be high and that the work in question can be more interesting precisely when they are.

This is a quietly important move. The earlier consensus, where it existed, treated the human-machine split as zero-sum. Every point of machine contribution was a point off human authorship; the more autonomous the system, the less of the work belonged to the artist. That framing made it nearly impossible to acquire ambitious generative work. A piece that was substantially machine-led had, by the math, a low human-authorship score, and a low score in any required column tended to disqualify the work.

The new rubric does not solve the problem. It moves it. A piece can now score five-five — high human, high machine — without contradiction. The committee then has to decide what that combination is worth, in deliberation, by argument. The rubric is honest about leaving the decision to the people in the room.

This, in practice, is what most institutional decisions actually are: an argument with a scoring sheet for cover. The rubric does not pretend otherwise. It admits that the question of how to weight machine contribution against human authorship is not a question a number can answer, and it gives the committee a vocabulary for the argument rather than a verdict.

A useful precedent: the way photography was eventually admitted to museum collections in the mid-twentieth century. The institutions then, also, had to invent a vocabulary that did not pre-exist. They had to decide whether the print was the work, or whether the negative was, or whether a series was inseparable. The decisions varied, the vocabulary was uneven, and most of the early framings were quietly revised. The rubric, when the rubric came, was a relief because it codified the argument the curators had been having anyway.

The leaked document is, in that sense, a relief. It does not pretend the new work is the old work. It does not pretend the old criteria still suffice. It admits that two questions are now in the room — how much person, how much machine — and that both deserve their own column.

What it does not do is tell the committee how to weight them. That is a public choice in waiting. It will be made, in fragments, over the next decade, by the works the museum chooses to acquire and the works it chooses to pass over. The rubric is the question. The collection that emerges from it is the answer.

The integrative read: the museum has done a quietly grown-up thing. It has admitted that the decision is hard and refused to reduce it to a number. It has given itself permission to argue. The argument will be uneven and occasionally embarrassing and sometimes wrong. It will also, because it is being had at all, produce a collection that knows what it thinks about what it bought.

That is more than most institutions managed in the photography era. It is enough.