SAG-AFTRA weighs a studio tax on digital performers
Unable to block synthetic performers outright, the union is considering a levy on studios that use them — shifting the argument from prohibition to pricing.
This is the shape most of these fights will settle into: not whether synthetic performers are allowed, but what they cost and who is compensated. Budget for that outcome rather than for a ban.
The most telling line in Variety’s reporting comes from Brendan Bradley, a member of SAG-AFTRA’s AI task force, describing a royalty on synthetic performers as “the best bad idea we’ve got in 2026.” That is a union conceding, on the record and before bargaining began, that it cannot stop the technology and is instead trying to put a price on it. For anyone who budgets union work, that concession matters more than any single clause, because it tells you which way the next decade of negotiation will run.
The mechanics under discussion were straightforward: studios that use a synthetic actor in place of a human could pay a royalty into a union fund. The idea did not come from nowhere. Variety notes that SAG-AFTRA had already won royalties for wholly synthetic performances in its deals with record labels and commercial producers, and that under the union commercials contract, if an actor is replaced by an AI performer, the equivalent pay has to go into the pension and health funds. The so-called “Tilly tax”, named after the digital creation Tilly Norwood that crystallised these fears, would extend a model that already exists in advertising into film and television.
The context was unusually calm for a contract cycle. The union agreed to bargain early, months ahead of the June 30 expiration, and the AMPTP had a new lead negotiator, Greg Hessinger, himself a former SAG leader. Not everyone inside the union was comfortable. Erik Passoja, a former co-chair of its L.A. New Technology Committee, called the levy a last resort: “I say no, no and triple no. But if it has to be, it should go into pension and health.” That split between principled refusal and pragmatic pricing is worth remembering, because it will resurface every time the terms are revisited.
In practice, the commercials precedent is the one that bites first. If you produce union spots, replacing a performer with a synthetic one does not remove the cost line; it redirects it into the funds. Producers who assumed a digital performer was a saving need to model it as a cost that still exists, plus whatever the synthetic work itself costs to build, clear, and maintain across a campaign. For film and episodic work, the ratified four-year deal covered in May framed the rule differently, permitting an AI performer only where it brings “significant additional value” over a live actor or that actor’s digital avatar. Whether any levy-style mechanism sits alongside that test is something to check in the contract language itself, which is why the clause-level read of the deal from The Hollywood Reporter is worth the time.
Our view is that pricing, not prohibition, is the durable settlement, and it is a healthier one for production than a ban would be. A ban invites workarounds and quiet non-union production. A price makes the choice explicit, documentable, and auditable, which is what legal and clearance teams actually need. The open question is enforcement: a levy depends on someone identifying when a synthetic performer has replaced a human one, and that line will blur as hybrid work, digital doubles, and partial replacements become routine.
The thing to watch is whether the definitions keep up. A royalty keyed to “wholly synthetic” performances is easy to administer; a royalty that has to handle a human performance modified by AI, or an avatar trained on several people, is not. That is where the next round of disputes will come from, and it is where production paperwork, from consent forms to shot-level records, will need to be more precise than it has ever been.