Actor and model — tasks, one by one
The unit of analysis is the task, not the job title. Each one below carries its direction, whether the judgement rests on evidence or on platform inference, the reasoning, and what it does not establish.
Every task on this page#
Principal screen performance
Still human-led✓ Evidence-backedPlaying speaking and featured roles in film, television and streaming.
The actors' union contract protects the performance itself. Under the 2023 agreement, a producer must obtain clear and conspicuous consent to create a performer's digital replica and pay when the replica performs scenes the performer would otherwise have done; under the 2026 agreement, producers agree to a principle strongly favoring human performances and must give prior written notice before using a synthetic performer in a human role. US projections expect actor employment to show little or no change from 2025 to 2035, while saying AI may replace some actors in certain subfields.
US union contracts for signatory producers and a projection; synthetic performers need notice and bargaining, not consent, and non-union work is not covered.
Background and extras work
Being augmented✓ Evidence-backedAppearing in crowds and background scenes.
Digital doubling of crowds is where replicas reach background actors, under limits. The 2023 agreement says a producer may not use a background actor's digital replica beyond the production they were hired for without consent and separate bargaining, but expressly leaves the tiling of background actors outside those rules.
Contract terms, not a count of background bookings; how often crowds are now doubled digitally is not measured here.
Fashion and commercial modelling
Automating✓ Evidence-backedPosing for campaigns, catalogues, e-commerce and advertisements.
Brands are generating model images. Mango created a campaign generated entirely by generative AI from real photos of the garments, and H&M published its first images featuring digital twins of models; Levi's announced tests of AI-generated models to supplement human models while saying it was not scaling back live shoots. US projections expect model employment to decline 2 percent from 2025 to 2035, saying AI that allows companies to reuse images of products and models may limit demand. People told AI-generated faces from real ones only 48.2% of the time.
Companies' own announcements of single campaigns and a pilot, a projection and a study of still faces; none counts model bookings.
Stage and live performance
Still human-led≈ Platform inferencePerforming live in theatre and events.
Live performance happens in the room, and no record here shows AI performing it. US projections for actors point to the television and film industry shifting from growth to profitability, adding only that AI may replace some actors in certain subfields.
The absence of evidence here is not evidence that live work is unaffected; no record measures theatre employment.
Likeness rights and consent
Still human-led✓ Evidence-backedControlling how one's face, body and performance are copied and used.
Law now protects performers' likeness. California and New York make contract clauses allowing a digital replica in place of in-person work unenforceable or void unless the performer had a lawyer or a union covering replicas; California makes producers liable for using a dead performer's replica without consent; and New York requires advertisers to conspicuously disclose a synthetic performer.
Two US states' law; other places differ, and these rules govern consent and disclosure, not whether replicas are used.