Max Einhorn spent a decade in film distribution before a revelation changed everything: What if you could create Colosseum gladiator scenes for a fraction of the millions of dollars it would typically cost? His company, Gennie, has proven this possible, producing the first documentary series using 100% generative AI reenactments that's now airing on Sky HISTORY.

Speaking with us at AI on the Lot for our Inside the AI Studioseries, Einhorn breaks down how generative tools are transforming documentary production by solving one of the industry's biggest budget challenges.

Watch the full interview here:

Einhorn's path to AI-powered documentaries started through his distribution work at FilmRise, where he specialized in using data analytics to find profitable content in library catalogs. This experience led him to co-produce international content, particularly true crime shows, where he witnessed producers' constant struggles firsthand.

"You never have enough budget. You can't quite get the right actors. You don't have enough for production design, props, location, shooting permits," Einhorn explains. "They never look good unless you have an enormous budget and nine times out of 10, a producer does not."

According to Einhorn, the breakthrough came around 2022 when synthetic avatar tools like Synthesia and Studio D-ID converged with language models and image generators like DALL-E and Midjourney.

Gennie's first major success is Killer Kings, an investigative documentary series about formidable rulers and the forces that drove them to kill. The show premiered on Sky HISTORY in June, featuring entirely AI-generated reenactments.

"We were really gonna go for it," Einhorn says about their approach. "We were not gonna cheat it in the same way that you might in a true crime show where you're showing somebody from behind, you're showing a hand, turning a knob, or you know, a knife on the counter in detail shots."

The production process relied heavily on historical research rather than traditional casting:

Einhorn's team learned quickly that current AI models aren't automatically broadcast-ready. When they delivered their first episode to OutPost Facilities at Pinewood Studios, 60% of clips were rejected for quality issues.

"It was a lot of softness. It was a lot of solve this and also lack, lack of detail," Einhorn recalls about the initial feedback.

The solution came through Topaz Starlight, an AI upscaling tool that addressed many of the video quality problems. However, the experience taught them important workflow adaptations:

Even with advanced AI tools, some scenes required creative problem-solving. Einhorn discovered that certain actions, like executions, wouldn't generate properly through text prompts alone.

"The executioner would, he would raise an ax over his head and then it would just stay up there and he wouldn't kind of know what to do," he explains.

His solution involved filming himself swinging a plastic ax in his backyard, then using video-to-video AI processing to create the final historical execution scene. According to Einhorn, viewers seeing the show will actually be watching his motion translated through AI into a historical context.

Einhorn believes 3D technology will play a crucial role in the evolution of AI-generated content, particularly for maintaining character and location consistency across different angles—something traditional productions take for granted.

He sees the tools becoming increasingly accessible as companies focus on expanding their user base and reducing technical barriers. This democratization could significantly impact how documentary producers approach historical content creation.

As generative AI tools continue improving, Einhorn's work with Gennie demonstrates how artificial intelligence can solve real production problems rather than simply creating flashy effects. The ability to generate historically accurate reenactments at a fraction of traditional costs could reshape how documentary producers approach storytelling, particularly for historical and true crime content.

For producers struggling with reenactment budgets, this approach offers a viable path forward—though success requires understanding both the capabilities and current limitations of AI video generation tools.