Albert Einstein advertises cryptocurrencies on TikTok. Judy Garland he recites lines never spoken in a fake documentary. Jimmy Stewart He lends his face to a political campaign he would never have supported. These aren't dystopian scenarios: this is what could happen (and is already partly happening) to deceased celebrities in the era of deepfakes if there aren't more agreements like the one just signed between CMG Worldwide e Lotus AI.
What is it? It's a pact that protects the digital identities of 25 historical icons, from Rosa Parks to Neil Armstrong, from unauthorized artificial reproductions. Because when you've been dead for decades, you can no longer defend yourself. But someone else can do it for you. And it must.
The system that protects celebrities from digital clones
The agreement between CMG Worldwide, a company that manages the image rights of historical celebrities, and Lotus AI, specialized in deepfake detection, provides Continuous monitoring, automatic detection and immediate removal of unauthorized content on all digital platforms. According to Variety, the Loti AI system will ensure that the celebrities managed by CMG are “represented with accuracy, consensus and respect”.
In addition to Einstein, Garland and Stewart, the protection covers Neil Armstrong, Harry Belafonte, Andre the Giant, Joe Louis, Rocky Marciano, the general George Patton, Burt Reynolds, Christopher Reeve, Sugar Ray Robinson, Ginger Rogers, Mickey Rooney, David Ruffin, Macho Man Randy Savage, Mark Twain, John Wayne e Raquel WelchA cast of celebrities who shaped the history of the twentieth century and who now risk becoming digital puppets in the hands of anyone with access to artificial intelligence software.
When dead celebrities become commodities
It was perfectly predictable, we have often talked about it on this site. The phenomenon of "digital resurrection" is not new. In 2019 Salvador Dalí “came back to life” to welcome visitors to the museum dedicated to him in Florida: an actor, artificial intelligence algorithms, and hundreds of scanned interviews have allowed the Spanish painter to be recreated with disturbing realism. Marilyn Monroe appeared on eight covers of CR China magazine in 2022, digitally wearing Balenciaga, Fendi, and Miu Miu. Sixty years after her death.
But it's not just the dead who are cloned. Eva Herzigova he sold his face to Hour One, a startup that creates digital clones for advertising campaigns. Lionel Messi lent his image to PepsiCo for deepfake commercials. Bruce Willis He appeared in a Turkish commercial without ever setting foot on the set. Tom Hanks had to publicly deny a deepfake video which showed him promoting a health insurance plan he had never sponsored. Taylor Swift she was the victim of pornographic deepfakes.
The threat is real and growing. According to DeloitteAI-based content fraud caused losses of more than $12 billion in the United States alone in 2023. Projections indicate it could reach $40 billion by 2027. A 2025 iProov study reveals that 72% of people fear being fooled by a deepfake on a daily basis.
Sora 2 and manipulation within everyone's reach
The urgency of the agreement between CMG and Loti AI is explained by the arrival of Gravel 2, the new video generator from OpenAI launched at the end of September 2025. The platform, according to the words of the CEO of Loti AI Luke Arrigoni, allows for “nearly unlimited content manipulation.” Sora 2 is described as “more physically accurate, realistic, and controllable” than previous versions in its ability to “modify objects and warp reality to successfully execute a text prompt.”
It's like Photoshop, but for videos. And you don't need any special technical skills. All you need is a photo and an audio track. As demonstrated by ByteDance's OmniHuman-1, a system capable of generating deepfakes of any length from minimal elements. An Einstein speech that never existed. A TED talk that never took place. All perfectly believable.
“The threat posed by AI reproductions not only risks damaging the reputation of celebrities, but it distorts history itself,” he said Joey Roesler, a lawyer for CMG Worldwide. Working with Loti will allow us to "protect their likenesses and preserve the integrity of the stories and legacies they have created."
Hollywood reacts, but it's already too late
Concerns about artificial intelligence were at the heart of the Writers Guild and Screen Actors Guild strikes in 2023. Clauses on deepfakes and the use of AI became critical elements of the contracts that ended the months of protests. But as the Center for Democracy and Technology in early 2024: “The battle against AI in Hollywood has just begun.”
According to the 2025 annual report of the Italian Data Protection Authority, 2024 was the year in which artificial intelligence had a direct impact on personal identity, with the GDPR acting as an “ethical-legal brake” pending the full entry into force of the European AI Act. Denmark prepares to become the first European country to systematically protect the body, features and voice with copyright.
CMG Worldwide's initiative comes at the right time. Or maybe it's already too late. Because if you can bring anyone back to life with a photo and an audio file, the question is no longer whether anyone will do it. The question is: who's going to stop those who are already doing it?
