Mark Wahlberg on AI in Film: Use It for the Stadium, Not Instead of Actors
Mark Wahlberg said he is open to an AI likeness continuing his work after he steps back. Speaking at TechCrunch Disrupt 2026, he said AI has a place in filmmaking as long as it is used the right way. TechCrunch
He spoke with TechCrunch Editor-in-Chief Connie Loizos, appearing alongside Bruce K. Lee in a session on entrepreneurship, investing and building startups. TechCrunch The conference ran October 13-15, 2026 at Moscone West in San Francisco.
His near-term example was the upcoming Pete Rose biopic, in which he stars as Rose. He said AI could potentially enhance aspects of the production, specifically the stadium setting.
The context for that example is environment work, not replacing performance. Stadium scenes require period accuracy, crowd density and steady lighting. That is visual-effects work where software already extends sets and fills seats.
He set a clear limit. Filmmakers should not enter an idea into ChatGPT, generate a script and make it without real actors. Storytelling, he said, should remain at the heart of filmmaking.
Loizos asked about the CAA Vault, the Creative Artists Agency facility that scans actors' bodies, faces, movements and voices to create AI replicas. Wahlberg said he had not visited the Vault but would be open to it.
His reason was family control. He said he was interested so his children or grandchildren could be in charge of his likeness and image. He added he would not oppose an AI version continuing his upcoming animated series Ted for decades.
The broader context here is a move from banning synthetic performers to managing them. A few years ago the debate was whether they should exist at all. Wahlberg accepts AI for environments, rejects scripts made without actors, and treats a lasting likeness as an estate asset. Use is allowed. Control stays with family.
In my view, the harder question is how that control works over time. That means storing original scans, tracking versions, keeping records of consent, and revoking permission when a use goes beyond scope. I have watched my own kids inherit passwords and photo libraries, and handoffs get messy fast. A face and voice carry higher stakes than a folder of photos.
Looking at what this enables, the Ted example is clear. An animated series already separates voice from physical presence, so a licensed voice and behavior model could sustain episodes without Wahlberg on a soundstage. The same holds for the biopic. Fill seats, fix period errors, extend the field. Actors keep acting and writers keep writing, without the cost of staging full scale. If storytelling stays central, as Wahlberg insists, the test is not whether AI was used. It is whether the story improves and whether the people behind the likenesses keep approval and pay. His rule is direct. Augment the world. Do not automate the teller.
