Hollywood Hasn’t Reckoned with the Impacts of Generative AI

When the company that runs the Las Vegas Sphere announced its plans for an adapted, immersive version of The Wizard of Oz using generative artificial intelligence (AI) in April, some veteran figures of classic film appreciation, including Turner Classic Movies host Ben Mankiewicz, offered the project their blessings. But the new attraction, which opened at the Sphere on August 28, has also prompted online backlash from critics who describe the project’s use of AI as “soulless” and “on the wrong side of film history.” 

This new “version” of the film increases the scope of what The Wizard of Oz fans see in the original product to fit the Sphere’s 160,000 square foot LED display, incorporating generative AI to build out new backgrounds. This 4DX project cost $100 million, uses 167,000 sound speakers, and cuts out around thirty minutes of the original film.

In the past year, AI has become increasingly prevalent in major film and television productions. Earlier this year, Natasha Lyonne sparked criticism after announcing that her upcoming directorial debut, Uncanny Valley, will incorporate generative AI with “traditional” filmmaking methods with an AI-focused production company she founded. In April, EDGLRD, a film studio created by Spring Breakers director Harmony Korine, recently made a deal with Runway, a film-focused AI research firm. AI has already made its way into high-profile releases: in the lead up to the 2025 Academy Awards, two Best Picture nominees, The Brutalist and Emilia Pérez, which both received Oscars, were both revealed to have used AI to augment actors’s voices.  

All of this comes on the heels of the most recent Screen Actors Guild and American Federation of Television and Radio Artists (SAG-AFTRA) strike, which ended in November 2023 after actors were guaranteed protections from their voices and images being replicated without their consent by AI, which is often used to create deepfake images and audio. Similarly, the SAG-AFTRA video game strike in July also centered around AI protections for use of performer likeness. 

But, according to many critics, seeing an untouchable Hollywood classic like The Wizard of Oz get the AI treatment raises deeper concerns for where the film industry may be headed next. The Wizard of Oz Las Vegas Sphere experience has used AI to create new performances with the likeness of deceased actors, including AI-generated appearances by Charley Grapewin, who plays Uncle Henry, in scenes he didn’t actually film.

In a CBS Sunday Morning special, Judy Garland’s daughter, actress Lorna Luft, says her mother would approve of the AI-generated material: “I think that she would really think, ‘How wonderful that this is happening.’ And I’m sure that she would love the audience reaction.” Garland’s other daughter, actress Liza Minnelli, openly supported the use of Garland’s voice in an AI app called Reader App last year, which would allow audiences to hear Garland read articles, essays, and books to them. But there’s no way to know for sure what Garland, her castmates, and the 1939 film’s original director, Victor Fleming, would actually think about having their likenesses generated by a technology that seemed unfathomable even ten years ago.

One of the SAG-AFTRA provisions secured through the 2023 strike requires a company to get consent from the performer (or their heirs in the event of their death) to use the performer’s likeness in any digital or AI replications. If there is no such person to provide consent, the producer or studio can then request the union grant them consent as a “discretionary decision” to use the performer’s image, SAG-AFTRA’s Chief Negotiator Duncan Crabtree-Ireland tells The Progressive. 

However, that consent standard isn’t required of non-union projects—which typically involve actors outside of SAG-AFTRA—prompting several industry groups, including the Television Academy and talent agency WME, to advocate for federal legislation that would protect all performers, regardless of whether they are unionized. “That is an area where we need public policy to catch up,” Crabtree-Ireland says. SAG-AFTRA is among those currently pushing for Congress to pass the No Fakes Act, which would protect the voice and image likeness of everyone—not just film industry performers—from unauthorized deepfakes created with generative AI. This would help regulate media by holding individuals and companies liable for their unauthorized digital replicas. Versions of the bill have been introduced in both the House of Representatives and the Senate this year with bipartisan backing. 

According to Crabtree-Ireland, the legislation has even been endorsed by tech companies, including Google and Open AI. “What it really implicates,” he says, “is not only the entertainment industry and the right of performers to control use of their image, but all of us, and how important it is that when we see someone’s face and voice saying something on social media or online or in any forum, we know if it’s really them.” 

Simultaneously, industry executives and film directors have been reluctant to address the environmental impacts of generative AI’s hardware, which requires substantial energy resources to run and keep cool, putting strain on the electric grid and municipal water supplies. As AI technology increasingly becomes a fixture in everyday life, Crabtree-Ireland says, it can be difficult to fully understand the waste it creates.

“I don’t think the industry’s taking it into account as much as it should,” he says. “It’s certainly something we’re interested in and concerned about. In fact, I was at a meeting with other members of the labor movement with one of the major AI companies, and I specifically brought up the topic of resource implications, [asking] what steps these companies are taking to either ensure that use is limited to what is actually necessary to achieve the outcome. I don’t think we’ve all heard a very satisfactory answer to that just yet.”

As some actors, directors, and filmmakers begin to embrace AI as an inevitable evolution of the film industry, the industry will have to reckon with deeper ethical questions. The backlash to The Wizard of Oz Sphere experience has shown that there will continue to be an unwavering crowd that opposes AI; moviegoing consumers now have the power to decide what they are willing to spend their money on. And that matters. But until the industry comes to terms with the unauthorized use of generative AI and its labor and environmental implications, the line between ethical and non-ethical use will only continue to stay blurry.