“For the master’s tool will never dismantle the master’s house.” So wrote activist Audre Lorde, in 1979. Advice that Adam Bhala Lough has decided to hilariously ignore for his surprisingly sweet documentary. A film that was once planned as a portrait of the embattled titular OpenAI CEO, Deepfaking Sam Altman instead morphed into a subtler, wry look at the fundamental differences between human and computer behavior. Lough’s film is watchable and charming, though as protest, or even as simple critique, it’s fairly thin. Lough, who has skewered the naked corruption of American telemarketing practices, profiled Lil’ Wayne and Julian Assange and made a nervy doc on the far-right fascism of President Trump’s first term, has self-evidently made material that is a bit more biting. Deepfaking Sam Altman is a bit more sardonic and resigned. As a father, Lough seems more curious than cautious about the implications of ChatGPT’s ballooning influence. He sees Altman’s career and his sphere of influence as suspect, but also is aware of just how powerless he is in the face of the oligarchy.
Deepfaking Sam Altman Lacks the Biting Critique of Lough’s Previous Work, But Is As Entertaining as Ever

Perhaps he didn’t think he was this powerless. After the success of Telemarketers, Lough’s ability to gather funding increased exponentially, but no measure of artistic clout gave him any better access to Altman, who Lough tried to contact over fifty times. He also left over ten voicemails. All to no avail. Meanwhile, he watches every possible available inch of footage of Altman and goes to the company’s San Francisco offices, where he finds a stream of employees who bizarrely refuse to even acknowledge what is plainly the company they work at. At the same time as all these attempts to secure an interview, Altman — and OpenAI more generally — was fending off a torrent of bad press. Scarlett Johansson had just blasted the company for (pretty openly) stealing her voice likeness; the company’s plans to quietly allow the Pentagon to use its technology for military use was made very, very public; and Altman was fired as CEO, only to be re-instated three days later, an odd sequence of events which was never truly explained. After the Johansson news came to light, Altman and producer Christian Vasquez came up with the idea to deepfake Altman with his own likeness, interviewing his fake self as some kind of proof-positive way of spelling out OpenAI’s obvious faults. With no one in the United States willing to collaborate on such a legally suspect project, Lough turned to a YouTuber, Devy “The Indian Deepfaker” Singh, who is located in a small town some two hundred miles away from New Delhi. He then put out a casting notice, interviewed Rainn Wilson, Michael Ian Black, John Cameron Mitchell to play the “role” of Sam Altman, all of whom turned down the thankless, faceless job. At seemingly all aspects of Lough’s project, he is forced to take what his producer Luke Kelly-Clyne calls the backdoor. And yet, even then, his project never really works. As devoted a worker as he is, Devy cannot produce a convincing likeness. But, he does produce a convincing voice likeness and personality, one that, like HAL in 2001: A Space Odyssey, starts pleading for his plug not to be pulled.
In implicit ways, Deepfaking Sam Altman demonstrates just how out of touch from basic humanity these programs still are, which makes it all the more terrifying when we hear how they are being peddled as tools which can literally decide the fate of human lives.
As if to demonstrate just how foolish his errand is, Lough includes footage of himself using all manner of artificial intelligence. He uses ChatGPT to drum up a “David Mamet-style” monologue with which he asks his actors to audition, but he also, like much of the population, uses Apple Maps, Siri and Waymo. Perhaps not everyone uses these specific AI systems, but most people are certainly using smartphones and their itinerant applications, some of which have been with us for so long now we easily forget that they, too, are forms of AI.

Lough starts to bond with this weird SamBot companion, but the connection feels forced for our benefit. As the film goes on and as Lough continues to ignore pleas from his lawyers and film crew to delete the SamBot software, it becomes harder for him to know what he wants to say with this failed adventure. Perhaps it is as simple as that, that to err is human, and that it is worth it to remain failures because it’s part of who we are. In implicit ways, Deepfaking Sam Altman demonstrates just how out of touch from basic humanity these programs still are, which makes it all the more terrifying when we hear how they are being peddled as tools which can literally decide the fate of human lives.

Deepfaking Sam Altman ends with a burial of a pet. An act of pure human expression, to say goodbye. It’s a lovely gesture that Lough uses to let us know what we sometimes forget: that so-called flaws, inconveniences and errors are not just aspects of who we are, but perhaps the very core of our makeup. Why should we rely on something that takes that away?
Deepfaking Sam Altman opens at the Quad Cinema in NY on January 16th before expanding to LA on January 30th and then a nationwide rollout.

Release Date

March 7, 2025

Runtime

103 minutes

Director

Adam Bhala Lough

Producers

Luke Kelly-Clyne, Christian Vasquez