Can you use AI in a film? What the Oscars, unions and the law say

Yes, you can use AI in a film. As of September 2026, no awards body, guild or U.S. law we checked bans AI tools in production or post. What the rules do regulate is three things: authorship (a person has to make the creative work), consent (performers have to agree before their likeness is replicated or materially changed) and disclosure (some awards bodies, distributors and laws want to know what AI did). Using AI the way post has always used tools, to paint out a rig or pull a matte, sits comfortably inside all three.

This post covers the Academy and the Golden Globes, the guilds, U.S. copyright, the state and EU laws on faces and voices, and a checklist for productions using AI in post. It’s a summary, not legal advice: check your contracts and ask counsel before you rely on any of it.

What do the Oscars and Golden Globes say about AI?

Both allow it, and both judge the human work. The Academy’s rules for the 99th Oscars (films released in 2026) say generative AI and other digital tools “neither help nor harm the chances of achieving a nomination.” Each branch judges the achievement by how far a human was at the heart of the creative authorship, and the Academy can ask a production for more information about how AI was used and who authored what (Academy rules, Rule Two). The rules announced on May 1, 2026 add two lines: acting nominations only go to roles performed by humans with their consent, and screenplays must be human-authored (Academy Press Office).

The Golden Globes go further on paperwork. Their rules for the 84th awards say AI doesn’t automatically disqualify a work as long as human creative direction, judgment and authorship stay primary, and they require every submission to disclose any generative AI used anywhere in the production, including any AI alteration of a credited performer’s likeness or voice. Technical or cosmetic changes to a performance can be acceptable if the performance stays the actor’s own (Golden Globes rules, section E).

For post, the practical reading is simple: cleanup, mattes and set extensions made with AI tools don’t hurt a film’s eligibility, but you may be asked to explain them.

What do SAG-AFTRA, the WGA and the DGA say about AI?

The guilds regulate AI rather than ban it, and the performers’ terms are the ones that touch post most directly.

SAG-AFTRA. A digital replica, as California law defines it, is a computer-generated, highly realistic electronic representation that is readily identifiable as a person’s voice or visual likeness. Under the 2023 TV/Theatrical agreement, creating and using a performer’s digital replica needs consent that is clear and conspicuous, with a reasonably specific description of the intended use. Digitally altering a recorded performance also needs consent, unless the performance stays substantially as scripted, performed or recorded. The agreement lists exceptions, among them cosmetics, wardrobe, continuity, visual effects, dialogue adjustments and lip adjustments for a foreign language (SAG-AFTRA 2023 summary; FAQ).

Members ratified a new agreement on June 4, 2026, in effect from July 1, 2026 to June 30, 2030 (SAG-AFTRA). Its summary makes three changes worth knowing in post (2026 summary, item 39):

• A replica built from the picture’s own photography, with no scan, is still an employment-based digital replica with all the protections. Face work made from your plates counts.
• “Scripted” now means only a script the performer had on or before their last day on the scene, so a later rewrite can’t be used to stretch the alteration exception.
• Using a digital replica to dub a performance into another language will need the performer’s consent for pictures starting principal photography on or after July 1, 2027. Producers also agree not to use a synthetic performer in a human role unless it brings significant additional value.

WGA. Under the 2023 Minimum Basic Agreement, AI can’t write or rewrite literary material, and AI-generated material isn’t source material for credit purposes. A writer can choose to use AI if the company agrees, but can’t be required to (2023 summary). The 2026 MBA, running May 2, 2026 to May 1, 2030, keeps those terms and adds notice to the Guild when companies license writers’ work to train commercial AI systems (2026 summary).

DGA. The agreement members ratified on June 26, 2026 renews the 2023 principle that directors’ work is performed by a person, and the Guild says it gives the director control over AI-generated footage (DGA, June 12, 2026; ratification).

Can a film made with AI be copyrighted?

Yes, the human-authored film can. The U.S. Copyright Office’s January 29, 2025 report on copyrightability concluded that using AI to assist, rather than stand in for, human creativity doesn’t affect copyright. Its example fits VFX exactly: “a film that includes AI-generated special effects or background artwork is copyrightable,” even if those effects and artwork aren’t protected on their own (Copyright Office, Part 2).

Two limits follow. Purely AI-generated material isn’t protected, and the Office found that “prompts alone do not provide sufficient human control” to make someone the author of the output. When you register a work with more than a trivial amount of AI-generated material, the Office’s guidance asks you to disclose it and describe the human contribution. The Supreme Court declined to hear Thaler v. Perlmutter on March 2, 2026, leaving the human-authorship requirement in place (docket 25-449).

For post, this favors the tool-shaped uses: an artist marks the region, chooses between versions and finishes the comp. That selection and finishing is the human authorship.

Which laws cover digital replicas and deepfakes?

Several, and they mostly concern faces, voices and the dead.

• California AB 2602 (signed September 17, 2024, Labor Code § 927) makes a contract clause allowing a digital replica of someone’s voice or likeness unenforceable for new performances from January 1, 2025, if it replaces work they would have done in person, lacks a reasonably specific description of the uses, and the person had neither a lawyer nor a union negotiating it (AB 2602).
• California AB 1836 (signed the same day, Civil Code § 3344.1) makes it a liability to use a deceased personality’s digital replica in an audiovisual work without consent from the estate, with exceptions such as news, parody, and documentary or biographical portrayals (AB 1836).
• New York signed two laws on December 11, 2025: one requires consent from heirs or executors to use a deceased person’s name, image or likeness commercially, and one requires advertisements to disclose AI-generated synthetic performers, in effect since June 2026 (signing; in effect).
• EU AI Act, Article 50, applies from August 2, 2026. Anyone deploying AI to generate or manipulate a deepfake must disclose it. For an evidently artistic, creative or fictional work, that duty is limited to disclosure “in an appropriate manner that does not hamper the display or enjoyment of the work” (Article 50; Commission FAQ).

Do you have to disclose AI in a film?

Sometimes, and more often than a year ago. The Golden Globes require it for every submission. The Academy can ask. In the EU, deepfake content needs a disclosure, light-touch for fiction. In New York, ads with synthetic performers need a label. We found no U.S. federal rule requiring a general AI credit on a film as of September 2026.

Distributors and insurers ask too. Netflix’s guidance for its productions says digital replica use may require documented talent consent, material alterations may too unless they stay consistent with traditional post work, and some uses need written approval from Netflix before you start (Netflix Knowledge Hub). On insurance, a Canadian entertainment broker told Insurance Business in July 2026 that errors and omissions (E&O) underwriters don’t cover AI use automatically and ask how the tool is used and what controls are in place (Insurance Business).

What does each rule mean for post-production?

As of September 2026. Terms change with each contract cycle, so check the current text.

How do you use AI in post without trouble? A checklist

Most of the risk sits in paperwork, not pixels. For a production using AI in post:

1. Log what touched which shot. Shot, tool, provider, date, what changed, who approved it. That record answers an awards committee, a distributor or an E&O underwriter.
2. Get performer consent in writing for any face, lip or likeness change, with a reasonably specific description of the use, before the shot goes out. Don’t rely on the alteration exceptions without counsel.
3. Keep human creative control visible. Versions reviewed, the chosen one signed off by the director or supervisor, the comp finished by an artist. This is your authorship record.
4. Check the provider’s terms on training and retention before footage leaves the building. Our client footage checklist lists the questions.
5. Prefer tools that change only what you mark. Plate-preserving edits are easier to explain than a regenerated frame; we compare the two in generative video editing vs plate-preserving VFX.
6. Tell your distributor and E&O broker what you used when they ask, and read your delivery requirements for AI clauses.
7. Plan disclosures early for awards submissions and EU releases, so they come from the log, not from memory.

Where nolanlabs fits

nolanlabs does VFX shot work from cleanup to performance assist, and its own rules line up with the checklist above. The filmmaker makes the creative calls: every result is a version a person reviews and chooses. Region jobs change only what you mark, and each run comes back as mattes, patches and layers an artist keeps refining. Footage is never used to train AI models, ours or our providers’, and we name the providers that process it.

For performance assist, the rule is fixed: “Faces and lips change only when the performer agreed to it, in writing. No voices, no synthetic performers, never a performance made from nothing.” We ask the production to confirm that consent in writing before any face or lip job. Hard head turns, hands across the face and heavy motion blur are still the hardest cases, and the production reviews every result. None of this replaces your contracts, clearances or disclosures, and no guild, academy or regulator has endorsed it. Those stay with you and your counsel. The reasoning is in our manifesto, and the wider picture of AI in post is in AI in post-production. Access is by invitation, starting with a demo: request a demo and bring a shot.

The short version: use AI where it saves hours, keep a person making the calls, get consent before anyone’s face changes, and write down what you did.

Questions

Can a film that used AI still win an Oscar?

Yes. The Academy’s rules for the 99th Oscars say generative AI and other digital tools neither help nor harm a film’s chances, and each branch weighs how far a human was at the heart of the creative authorship. Acting nominations are limited to roles performed by humans with their consent, and screenplays must be human-authored. The Academy can ask a production for more information about how AI was used.

Do you need an actor’s consent to change their face or lips with AI?

Often, yes, and it is always safer to get it in writing. Under SAG-AFTRA’s TV/Theatrical terms, creating or using a performer’s digital replica needs their consent, and so does digitally altering a recorded performance, unless it stays substantially as scripted and performed or falls under listed exceptions such as cosmetics, continuity and visual effects. California and New York law add their own rules for digital replicas, so check the contract and ask counsel before any face or lip work.

Can a movie made with AI be copyrighted in the US?

Yes, the human-authored parts can. The U.S. Copyright Office said in January 2025 that using AI as an assistive tool doesn’t affect copyright, and that a film with AI-generated effects is copyrightable as a whole even if the AI material alone isn’t. Purely AI-generated material, or output where a person only wrote prompts, isn’t protected.

Do you have to disclose AI use in a film?

Sometimes. The Golden Globes require every submission to describe any generative AI used in the production, Netflix asks partners to get written approval for certain uses, and the EU AI Act requires deepfake content to be disclosed from August 2, 2026, in a light form for fictional works. We found no U.S. federal rule requiring a general AI credit on a film as of September 2026, but distributors, insurers and festivals can ask, so keep a record.

Sources

1. 99th Academy Awards Rules — Rule Two: Eligibility, paragraph 7 (May 2026) — Academy of Motion Picture Arts and Sciences — accessed 2026-09-27
2. Awards rules and campaign promotional regulations approved for 99th Oscars (May 1, 2026) — Academy Press Office — accessed 2026-09-27
3. 84th Annual Golden Globes Eligibility and Consideration Rules, section E: Artificial Intelligence (May 2026) — Golden Globes — accessed 2026-09-27
4. 2023 TV/Theatrical Contracts: Summary of Tentative Agreement, section II: Artificial Intelligence — SAG-AFTRA — accessed 2026-09-27
5. Artificial Intelligence Resources (2023 TV/Theatrical FAQ) — SAG-AFTRA — accessed 2026-09-27
6. SAG-AFTRA members approve 2026 TV/Theatrical Contracts tentative agreement (June 4, 2026) — SAG-AFTRA — accessed 2026-09-27
7. 2026 TV/Theatrical Contracts: Summary of Tentative Agreement, item 39: Artificial Intelligence (May 11, 2026) — SAG-AFTRA — accessed 2026-09-27
8. Summary of the 2023 WGA MBA — Writers Guild of America — accessed 2026-09-27
9. Summary of the 2026 WGA MBA — Writers Guild of America West — accessed 2026-09-27
10. DGA National Board unanimously approves tentative new agreement (June 12, 2026) — Directors Guild of America — accessed 2026-09-27
11. DGA membership ratifies new contract by overwhelming margin (June 26, 2026) — Directors Guild of America — accessed 2026-09-27
12. Copyright and Artificial Intelligence, Part 2: Copyrightability (January 29, 2025) — U.S. Copyright Office — accessed 2026-09-27
13. Docket No. 25-449, Thaler v. Perlmutter (petition denied March 2, 2026) — Supreme Court of the United States — accessed 2026-09-27
14. AB 2602 (Chapter 259, Statutes of 2024), Labor Code § 927 — California Legislative Information — accessed 2026-09-27
15. AB 1836 (Chapter 258, Statutes of 2024), Civil Code § 3344.1 — California Legislative Information — accessed 2026-09-27
16. Governor Hochul signs legislation to protect consumers and boost AI transparency in the film industry (December 11, 2025) — New York State — accessed 2026-09-27
17. First-in-the-nation law requiring disclosure when advertisements include AI-generated synthetic performers is in effect (June 9, 2026) — New York State — accessed 2026-09-27
18. Article 50: Transparency obligations for providers and deployers of certain AI systems — EU AI Act (Regulation (EU) 2024/1689) — accessed 2026-09-27
19. Transparency obligations under Article 50 of the AI Act, FAQ (July 24, 2026) — European Commission — accessed 2026-09-27
20. Generative Workflows, Applications, and Use Cases — Netflix Studio Partner Knowledge Hub — accessed 2026-09-27
21. AI hasn't created new risks in film production, just new ways into old ones, BFL's Lindner says (July 15, 2026) — Insurance Business — accessed 2026-09-27

Written by the nolanlabs team. nolanlabs does AI shot work for post-production; product names mentioned belong to their owners. Tool details were accurate on the date shown above — check the vendor’s documentation before relying on them.