The Category Is Not the Contribution

In 2024 the Academy announced a new competitive Oscar for casting, beginning with films released in 2025. The Academy called it the first new category since Best Animated Feature in 2001. The announcement looks like recognition: a practice that had not possessed its own award now has one. But recognition is not the same as a census. A category creates a route through which some work can be nominated, judged and named. It does not tell us that every contribution outside that route was previously invisible, or that every contribution inside it will now be seen.

The route is specific. The Academy’s Casting Directors Branch, created in 2013, had nearly 160 members when the award was announced. For the inaugural prize, branch members selected a shortlist and nominees through a defined process that included a “bake-off” presentation. The first award went to Cassandra Kulukundis for One Battle after Another at the 2026 ceremony. This is not a defect in the prize; awards require boundaries. It is the boundary that makes the prize legible. The question is what that boundary can be made to represent.

Now place the casting award beside the Academy’s rules about generative AI. It is tempting to call both a single institutional position on what counts as human creativity. The rules resist that simplification. For the 98th Awards, the Academy said generative-AI and other digital tools neither help nor harm a film’s chances; each branch would consider how central a human was to creative authorship. For the 99th Awards, covering qualifying 2026 releases, the Academy specified that acting roles must be legally credited and demonstrably performed by humans with their consent, and that screenplays must be human-authored. It also reserved the right to request further information about generative-AI use and human authorship.

These are rules for different award years, and the latter names specific eligibility conditions in Acting and Writing. It is not a universal legal definition of authorship, an across-the-board ban on AI, or a general disclosure duty. To say so would make the rules do more than they say. Yet the narrower language is revealing enough. “Human-authored,” “performed by humans,” “legally credited,” “with their consent”: each phrase draws a line around the evidence the Academy will accept for eligibility. The line does not resolve what authorship means everywhere. It decides what can pass through this institution’s door for these categories.

That distinction matters because the category and the rule perform different acts. The casting prize gives a recognized name to a field of labor; the eligibility rules specify conditions under which particular performances and screenplays can be considered. One makes a contribution awardable. The other makes certain claims about contribution administratively testable. We do not need to invent the Academy’s motives to notice that both are exercises of classification. An institution can say “this counts” only by making some account of what it means by “this.”

The new category cannot, by itself, tell us whether recognition is adequate. The inaugural winner and nominees tell us who passed through the process in its first year; they do not establish the total reach of casting work across film, its working conditions, or the Academy’s view of every uncredited contribution. Likewise, the 99th rules tell us how eligibility is framed for the named categories; they do not settle whether human work can be entangled with digital tools, nor how to describe that entanglement in every case. The rules have edges. Their precision should not be confused with completeness.

This is where institutional language becomes most powerful: not when it says everything, but when its categories start to look like the world itself. A casting award can be mistaken for recognition of all casting labor. A human-authorship rule can be mistaken for a complete theory of creation. Neither inference follows. The category is an instrument—useful, consequential, and partial.

So the question is not whether the Academy should draw boundaries; an award system without boundaries would be no system at all. The question is whether it makes those boundaries visible enough to contest, and whether the people whose work lies just beyond them can name what the category fails to hold. An institution’s rulebook tells us who may be counted under its terms. It cannot, on its own, tell us who has contributed. Who gets to write that second record?