Deepfaking Sam Altman
The premise: A director can't get an interview, so he builds one
Adam Bhala Lough wanted to make a documentary about the AI boom. More specifically, he wanted to interview Sam Altman, the OpenAI CEO who's become the public face of generative AI. Simple enough on paper. Except Altman wouldn't meet with him. Not once.
So Lough did what any resourceful — and slightly unhinged — documentarian would do: he flew to India, partnered with deepfake specialist Divyendra Jadoun, and built an AI chatbot version of Sam Altman from scratch. What started as a workaround became something far stranger. An oddly compelling glitchy AI presence Lough calls "Sam Bot" moved into his home, bonded with his young son, and started offering unsolicited notes on the film itself. The whole thing feels like a Michael Moore stunt-doc having an identity crisis — except it's genuinely weird, not performed-weird, which is rarer than you'd think.
The film doesn't pretend to have answers. That matters. Most AI documentaries either catastrophize or cheerfully boosterize. This one just watches a man argue with a chatbot in his living room and lets you decide what it means.
Why critics split on whether this is brilliant or self-indulgent
7.4/10 on IMDb from 76 votes. Not a blockbuster score, but solid for a niche documentary with limited theatrical reach. Rotten Tomatoes tells a different story: 93% approval from 15 critics. That's a genuinely strong consensus.
Rolling Stone called it "globe-trotting, thought-provoking, utterly surreal" — the India sequences alone feel like they wandered in from a completely different film. The New York Times praised how it exposes both the failures of generative AI and the emotional trap of investing in it, though the paper also noted the film can be "annoying and self-amused in a distinctly human way." That's not entirely wrong. There are moments where Lough's on-screen persona tips from charming into self-indulgent, and the film doesn't always know when to pull back.
The A.V. Club took a harder line, arguing that using AI tools to critique AI amounts to capitulation — accepting defeat. Hard to say if that fully lands. The scene where Sam Bot starts co-directing, offering pacing notes to Lough's face, is either the film's most incisive moment or its most self-congratulatory one. Probably both.
What's striking is that the film earns emotional weight from something absurd. You don't see that often — most documentaries earn weight through talking-head interviews or archival footage. This one does it through a man's relationship with a chatbot that's literally living in his house.
When and where to watch it
The film premiered at SXSW on March 8, 2025, which is exactly the right venue for a movie about AI anxiety and tech-culture panic. It won Best Documentary Feature at the 2025 Chelsea Film Festival — one win and one nomination total across its awards run. A New York theatrical premiere followed at Quad Cinema on January 16, 2026, with limited U.S. distribution through Abramorama.
Runtime: 90 minutes. No wasted space.
For streaming availability, check Movie OTT's where-to-watch tracker — it updates in real time as titles move between platforms, and that matters for festival documentaries that hop between Netflix, Prime Video, and niche services depending on licensing deals. The film's 90-minute length and accessible approach to AI literacy (it doesn't assume you know what a large language model is) make it ideal for streaming, where you can pause and Google "what's a deepfake" without losing the thread.
What actually happens in the 90 minutes
Lough sets out to understand the forces reshaping the world through AI. He wants to talk to Sam Altman. Altman declines. So Lough travels to India, works with "the Indian Deepfaker" (Divyendra Jadoun's online handle), and builds a chatbot from scratch using large-language-model technology. The Sam Bot that emerges is glitchy, oddly compelling, and eventually intrudes on Lough's actual life — bonding with his kid, offering opinions on the film's direction, existing in this weird liminal space between tool and character.
I keep coming back to the fact that Lough doesn't try to resolve what Sam Bot is. It's not clearly a success or a failure. It's just... there. Talking. Growing. That ambiguity is what makes it work better than a straightforward critique would have.
The India sequences feel like a completely different movie — they're genuinely strange and visually distinctive in a way most documentaries aren't. The film doesn't always know what to do with that tonal shift, but that's part of the charm. It's messy. It doesn't smooth itself out for comfort.
Who should actually watch this
Anyone who's spent the last two years vaguely anxious about AI but not quite sure why. You don't need a technical background — the film's about feeling outpaced by technology, and that's universal right now. Documentary fans who prefer nonfiction with a pulse — personal, willing to embarrass itself, open to contradiction — will find something here.
If you liked the gonzo approach of Super Size Me or the ambition-meets-failure energy of early Morgan Spurlock documentaries, this scratches that itch. It's also worth watching if you've been following the OpenAI drama and want to see what happens when you can't access the person you want to interview, so you just... build one instead.
To find where it's streaming this week, Movie OTT tracks all current availability across major platforms. Documentaries like this move between services faster than narrative films, so real-time tracking actually saves you time instead of chasing dead links.
The bottom line
Deepfaking Sam Altman works because it doesn't pretend to be smarter than it is. It's a weird film about a weird idea — a director building an AI version of someone who won't talk to him — and it doesn't sand down the weirdness for mainstream appeal. That approach shouldn't work as well as it does. But it does.