Tuesday, November 5, 2024

Scarlett Johansson’s AI Legal Threat Sets Stage for Actors’ Battle With Tech Giants

Must read

Artists were first to sue. Then authors hit generative artificial intelligence companies with a volley of lawsuits, followed by publications. As battle lines over the use of AI tools in Hollywood are being drawn, actors may be the next group of creators to open another front in what could be an industry-defining legal battle against AI firms over the use of copyrighted works and personal data to power their human-mimicking chatbots.

On Monday, Scarlett Johansson threatened legal action against OpenAI for allegedly copying and imitating her voice after she refused to license it to the company. According to the actress, OpenAI asked her to be one of the voices called “Sky” for its newest AI system. She declined, though she said that didn’t stop chief executive Sam Altman.

“When I heard the released demo, I was shocked, angered and in disbelief that Mr. Altman would pursue a voice that sounded so eerily similar to mine that my closest friends and news outlets could not tell the difference,” she wrote in a statement.

Johansson said that the similarity was intentional, pointing to Altman tweeting “Her” — a reference to her role as an AI assistant who forms an intimate relationship with a human in Her. She’s hired legal counsel, who wrote two letters to OpenAI, directing them to detail the process by which they created the “Sky” voice. Both of the letters referenced in Johansson’s statement (she wrote them herself) were sent after the Altman-led firm rolled out its demo and advanced multiple potential legal claims, a person familiar with the situation tells The Hollywood Reporter.

“They wouldn’t have done this if not for the letters,” this source says. “This wasn’t just a ‘What’s going on over there?’ [letter]. This was much more aggressive and forceful.”

OpenAI dropped “Sky,” but the specter of litigation looms over the company embattled with legal issues over the foundation of its technology.

The legal threat follows the filing of a proposed class action in New York federal court against Berkeley-based AI startup LOVO accusing the company of stealing and profiting off of the voices of actors, as well as those of A-list talent such as Johansson, Ariana Grande and Conan O’Brien. It’s believed to the first lawsuit against an AI firm over the use of likenesses to train an AI system and marks a growing rift between creators and companies alleged to indiscriminately hoover troves of copyrighted works and data to fuel their technology.

For SAG-AFTRA, OpenAI’s blunder couldn’t have come at a better time. There has been a surge in AI services that allow users to replicate members’ likenesses without consent or compensation. The union has blitzed legislators in advocating for a federal right of publicity law in the absence of federal laws covering the use of AI to imitate actors’ likenesses. A patchwork of state right of publicity laws has filled the void, but OpenAI’s alleged theft of Johansson’s voice underscores limitations of the current legal landscape.

“The incident highlights the importance of protecting your voice in an age of AI. It is no longer science fiction to easily clone a voice, it is science fact,” says a SAG-AFTRA spokesperson. “Whether you are a professional performer looking to protect your career or an individual looking to protect the words ascribed to you, the need for federal protection is now.”

SAG-AFTRA has played a part in the introduction in Congress of three bills, including two that would create federal voice and likeness rights and another that would provide criminalize nonconsensual deepfaked sexual imagery. That’s not all, with efforts from the union to discourage the use of AI in the production pipeline. New York is considering legislation that would bar tax credits when AI is used to displace labor. In March, first-of-its-kind legislation was signed into law by Tennessee barring the use of AI to mimic a person’s voice, with violations classified as a criminal offense.

The union maintains that training AI systems on members’ likenesses without consent is a violation of their rights. A court will likely decide the issue.

With an A-list celebrity publicly clashing with OpenAI, the incident demonstrates a growing rift between tech firms encroaching into Hollywood and creators who fear being displaced by the tools they may have inadvertently helped create. There’s growing mistrust of AI companies, with many believing that the Altman-led firm isn’t operating in good faith.

OpenAI CTO Mira Murati said in a March livestream that the voice assistant wasn’t meant to sound like the actress. And before Johansson published her version of events, the company said in a blog post that it “believed that AI voices should not deliberately mimic a celebrity’s distinct voice” while omitting information that the actress refused to license her voice.

OpenAI executives have repeatedly declined to answer questions on whether its text-to-video tool Sora was trained on YouTube videos, with Murati saying the company used “publicly available data and licensed data.” It also no longer discloses the materials used to train its AI system, attributing the decision to maintaining a competitive advantage over other companies. The firm has been sued by several authors accusing it of using their copyrighted books, the majority of which were downloaded from shadow library sites.

There’s a distinctly Silicon Valley ethos in OpenAI’s actions: Ask forgiveness, not permission.

“If they can do this to Ms. Johansson, imagine what they’ll do to the 23-year-old screenwriter who’s just getting started,” says Justin Nelson, a lawyer for authors suing OpenAI and Microsoft, “or some actor who just moved to Hollywood and doesn’t have near the resume” as the actress.

Though Johansson is unlikely to sue now that OpenAI dropped “Sky,” essentially granting her an injunction without having to file a lawsuit, OpenAI faced an uphill battle in court if it was met with litigation, according to legal experts consulted by THR.

A lawsuit from Bette Midler against Ford over a series of commercials called “The Yuppie Campaign” in which the company used an impersonator of the singer to imitate her voice may be instructive. Like Johansson, Midler was asked to sing for the adverts but refused. Ford subsequently hired a voice-impersonator to sing one of her songs in the commercial. After it aired, she was told by “a number of people” that it “sounded exactly” like her, according to court filings.

After the federal judge overseeing the case granted summary judgment to Ford, finding no rights preventing use of her voice, Midler appealed to the 9th U.S. Circuit Court of Appeals. That court ultimately found that the singer’s voice is distinctive to her identity, which Ford profited off of. The ruling establishes rights to uncopyrightable identifiers, such as a voice, when an individual who’s famous for that feature is involved.

Essential to the court’s order was Ford’s motivations for using a Midler impersonator. The questions that were asked by the justices included why the company asked Midler to sing if her voice wasn’t of value and why the impersonator was instructed to imitate the singer.

Intellectual property lawyer Purvi Albers says OpenAI’s solicitation of Johansson’s services is vital to whether the company violated her publicity rights. “It’s clear that was the voice they were going for,” she adds. “They wanted to piggyback off of her husky voice.”

The upshot: It may not matter if Johansson’s voice wasn’t used to train “Sky” as long as the goal was to capture her performance in the Spike Jonze film.

A lawsuit could’ve advanced right of publicity, right of privacy and possibly a federal trademark claim over potential confusion of consumers believing she’s associated with OpenAI. A person familiar with the situation said that there were “certainly references beyond simple right of publicity” claims in the letter.

“Sky” may also have implicated the rights of Annapurna, which produced Her, if Altman was looking to replicate Johansson’s performance in the film.

In the entertainment industry, the specter of AI casts a daunting shadow. A study surveying 300 leaders across Hollywood issued in January reported that three-fourths of respondents indicated that AI tools supported the elimination, reduction or consolidation of jobs at their companies. Over the next three years, it is estimated that nearly 204,000 positions will be adversely affected.

Paul Skye Lehrman, who has over a decade of experience as a voiceover artist and is suing an AI startup, said he’s gotten roughly 50 percent less work since last year. He stressed that the issue is not only that he’s getting less job opportunities but the “degradation of my reputation.”

“My voice is literally saying things I wouldn’t say with brands I wouldn’t work with in places I wouldn’t want to be placed,” he added.

Latest article