A Filmmaker Created a Sam Altman Deepfake and Developed an Unforeseen Connection

Director Adam Bhala Lough didn’t initially aim to create a documentary focused on a digital version of Sam Altman.

However, after approximately 100 days of attempting to secure an interview with the OpenAI CEO through texts and emails—without a response, as he claims, and with investors pressing him to deliver on his initial idea—Lough found himself at a breaking point.

He had explored nearly every angle. “Once I reached that stage, I conceded defeat and decided to gate-crash OpenAI,” he shares. Though he had used a similar approach in his Emmy-nominated 2023 documentary Telemarketers, which explores widespread corruption in the telemarketing sector, it wasn’t a directorial method he felt particularly comfortable with. “It was like a fortress. I managed to slip through the entry, but security promptly apprehended me and physically removed me.”

Thus begins Deepfaking Sam Altman, Lough’s exploration of AI’s impact on society alongside his pursuit to converse with its architect. When his initial strategy fell apart, he found inspiration in Altman himself. “The Scarlett Johansson controversy surfaced,” he notes. In 2024, the actress publicly criticized OpenAI for allegedly mimicking her voice in its new AI voice assistant, Sky. “That’s when the idea to create the deepfake struck me.” (In a May 2024 statement, Altman expressed regret to Johansson, clarifying that Sky’s voice was “never meant to resemble” hers.)

What begins as a straightforward voice clone expands into a full-fledged deepfake of Altman named Sam Bot, which Lough travels to India to develop. With Lough at the helm, however, events take unexpected turns. Without revealing too much, Sam Bot eventually evolves into a distinct entity, leading the film into even more curious—and insightful—territory. “There are parallels between this movie and Terminator 2: Judgement Day, but without any of the violence,” he explains. Lough grew up during what he refers to as the “AI 1.0 era,” and his fascination with James Cameron’s Terminator 2 significantly shaped his artistic vision.

Deepfaking Sam Altman, which draws some inspiration from a New York Magazine piece portraying Sam Altman as the Oppenheimer of our time, includes insights from former OpenAI safety engineer Heidy Khlaaf, who informs Lough, “We’re beginning to see OpenAI explore military applications, and I can’t fathom Dall-E and ChatGPT being utilized for military purposes. That truly concerns me, especially considering how inaccurate those systems can be.”