Artificial intelligence (AI) legal developments are occurring at breakneck speed, so this article provides a brief update on recent AI legal developments that have taken place in 2025 with regard to the entertainment industry.

Right of Publicity

California

Effective as of January 1, 2025, California prohibits contracts that purport to permit the creation of digital replicas of a person unless a number of easy-to-meet exceptions apply, such as if the person is represented by a lawyer or use of the digital replica does not replace work that the person would otherwise do. Another exception is if the person is a member of SAG-AFTRA, because the SAG-AFTRA Basic Agreement provides much better protection.

California law provides the heirs of “deceased personalities” (persons who are widely known at the time of, or because of, their death) with right of publicity protection for 70 years after their death. However, there was a large statutory loophole that permitted use of the deceased personality’s voice or likeness in connection with “entertainment, or a dramatic, literary, or musical work,” and this exception was repealed as of January 1, 2025 with respect to digital replicas.

New York

In June, New York passed legislation requiring any advertisement that includes an image of a human created by AI (“synthetic performer”) to conspicuously disclose that fact unless (a) the image “is recognizable as any identifiable natural performer” (in which case that person presumably gave consent or has a right of publicity claim) or (b) the advertisement is for an expressive work (like a film) in which the synthetic performer appears.

Federal Laws

Given the wildly inconsistent rulings on the right of publicity at the state level, there is a desperate need for uniform federal legislation on this issue, as there is for copyright and trademark. The very first federal legislation directly addressing content created by AI was enacted in May, called the TAKE IT DOWN Act. The good news is that the Act creates criminal and civil penalties for using AI to create a look-alike version of someone without their consent, but the bad news is that it only applies if the person is shown as nude. This Act would have been the perfect opportunity to include ALL AI created look-alikes without consent, but alas, the federal bill that would do that; the NO FAKES Act, is still languishing in Congress and does not look likely to pass anytime soon.

Copyright

Uploading

There has been an avalanche of litigation by copyrighted content owners (e.g., text, photos, films, and music) against AI companies for copyright infringement for uploading the copyrighted content. The outcome of these cases will depend on the fair use defense, and in February, the case of Thomson Reuters v. ROSS Intelligence held that the fair use defense didn’t apply to uploading to train AI, at least when the uploading was done to compete with the uploaded content. The court relied in part on the 2023 Supreme Court case in Andy Warhol Foundation for the Visual Arts, Inc. v. Goldsmith, which essentially held that it is not fair use if the owner of the copyright would reasonably expect to be paid for the copying/use.

However, in June, two other courts (in Bartz v. Anthropic and Kadrey v. Meta) held that uploading books was fair use, at least as long as the output (a) did not infringe on the uploaded content and (b) did not hurt the market for the uploaded content, looking only to the prior market of sales of the books to consumers and not the market of licensing the books to AI companies. The courts predictably relied on Authors Guild v. Google, a 2015 Second Circuit case that permitted Google to copy the full text of books for purposes of permitting searches that resulted in showing just snippets from the books.

And in July, President Trump gave his verdict, which is that uploading is always fair use, a decision that might have an impact on how the Supreme Court rules on this issue in the future.

Copyright Protection

In March, the U.S. Court of Appeals for the DC Circuit held that AI-generated content does not qualify for copyright protection, following the position taken by other courts and the U.S. Copyright Office.

U.S. Copyright Office Report

In May, the U.S. Copyright Office released a “pre-publication” report on AI that concluded that (a) the mere fact of uploading by AI could be copyright infringement of the uploaded works and (b) AI output could be copyright infringement of the uploaded works even if the output were modified versions of the uploaded works. However, the very next day the head of the office was fired, throwing into doubt whether this “pre-publication” report will ever become final.

Disney and Universal v. Midjourney.

In June, Disney and Universal sued Midjourney for both uploading and downloading, since there was rampant and flagrant use of their characters by Midjourney users, which was encouraged by Midjourney’s website by hosting multiple examples of the characters being used in various settings. My bet is on Disney and Universal prevailing on this one under a “Napster” theory, since if they don’t, copyright protection could practically be considered gone.

Defamation

Given AI’s tendency to “hallucinate,” it is not surprising that AI has reported nasty untrue facts regarding real people, which has led to defamation claims against AI companies. An interesting aspect of this issue is that if the matter involves a public figure or a matter of public interest, a defense to a defamation claim is that the person making the statement actually believed it to be true (so no “actual malice”). In May, the case of Walters vs. OpenAI held that false derogatory content about the plaintiff produced by ChatGPT did not constitute defamation because the plaintiff was a public figure, and ChatGPT was incapable of “believing” anything it produced. The court held that even if the plaintiff were not a public figure, no reasonable person should believe ChatGPT derogatory content given that it provides warnings that it hallucinates.

Unhappy SAG-AFTRA

In May, SAG-AFTRA filed an unfair business practice claim with the National Labor Relations Board (NLRB) against the producer of the videogame Fortnite for alleged failure to bargain in good faith prior to using AI to recreate Darth Vader’s voice with the consent of the actor’s estate. The matter was settled when the parent company entered into a new agreement with SAG-AFTRA in July.

Really?

Almost exactly two years ago, a lawyer was sanctioned for filing a brief written by AI that was full of hallucinated case citations, and I was sure this would be a one-time event given the publicity surrounding the matter. However, such sanctions started occurring monthly, then weekly, and now daily! So, there are a lot of lawyers out there that put a tad too much trust in AI and continuing to rely on it without proper oversight.

This article was originally published in the November 11th issue of Primerus Weekly, in their Off the Record section.

Copyright Protection & Enforcement

Resources

Contact

Connect