Tom Cruise vs. Brad Pitt AI Fight Scene Sparks Debate Over Authenticity

An AI-generated video of Hollywood superstars Tom Cruise and Brad Pitt fighting went viral last week, as it reignited fears across the industry about the rapidly evolving technology and how it could potentially impact the movie business. Deadpool writer Rhett Reese admitted to being “shook” by the quality of the scene and declared that “it’s likely over” for traditional filmmakers. However, the clip has now come under fire for its questionable authenticity, as one digital creative has uncovered possible evidence of digital deception.

The video surfaced on X/Twitter, courtesy of Irish filmmaker Ruairi Robinson, who claimed that the Cruise vs. Pitt showdown was made by entering a “two-line prompt” into ByteDance’s newly launched Seedance 2.0 generative AI tool. The original post has garnered over 1.8 million views since it was published on Feb. 11, but has drawn mixed reactions, with some branding it “AI slop” and others, like Reese, viewing it as a disruptive force that threatens Hollywood as more studio leaders adopt these tools in a bid to cut costs.

However, all is not what it seems when it comes to Cruise and Pitt’s rooftop brawl – at least, that’s according to Aron Peterson, known online as Shokunin Studio. The media production professional led the charge in debunking the authenticity of the Cruise vs. Pitt fight scene as a “pure” AI creation in an eye-opening blog post, titled “Is it all over for filmmakers?” Peterson identified a major red flag as he found green-screen footage on Seedance’s own website featuring stuntmen performing the exact same fight choreography as the two Hollywood heavyweights. He argued that the viral clip used a video-to-video workflow – essentially swapping the stuntmen’s faces for Cruise and Pitt – rather than generating the scene from scratch.

Peterson posted a video on YouTube (shared above) showing the green screen footage used for a different Seedance demo alongside the AI fight between Cruise and Pitt. The evidence presented in the exposé drastically changes the outlook on AI, as it suggests that while the technology is evolving, such high-quality results still require the expense of a studio, stuntmen, a choreographer, lighting crew and a cameraman, which Peterson estimates to “cost a couple of grand a day on the low end” for roughly 20 seconds of footage.

In the blog post detailing his findings, Peterson shared further evidence of the digital deception in the video, as he pointed out that the handheld “shaky cam” movement is something current AI models struggle to simulate realistically and, when inspected closely, some of the punches in the Cruise vs. Pitt video never actually land, which is a common technique for real-life stuntmen to avoid injury, but in a purely AI-generated video, he argued, that “a hit should always be a very clearly visible hit.” This prompted the skeptic to seek clarity and ask for transparency from the tech giant behind the Seedance model and its user:

“So the ball is in your court Seedance and Ruairi Robinson. Was the input really just a 2 line prompt or was it actually 2 lines, green screen video footage, and face references too? The evidence appears to show that stuntmen were filmed from several angles, that a clip had to be generated for every angle, and then finally all clips were stitched together for marketing.”

This is a developing story…