The documentary landscape surrounding artificial intelligence has become deeply polarised, with two recent films offering starkly different interpretations of technology that promises both extraordinary benefit and catastrophic risk. The division they illustrate runs far deeper than cinematic choice; it reflects a fundamental disagreement about how humanity should approach one of the most consequential inventions in history.
The AI Doc: Or How I Became An Apocaloptimist, co-directed by Daniel Roher and Charlie Tyrell, attempts something ambitious: to navigate the conceptual chasm between those who believe artificial intelligence represents an existential threat to humanity and those who view it as the key to human flourishing.The film follows a father-to-be trying to make sense of the current AI frenzy, from hype and investment to real fears about what happens when the tools outpace the guardrails.
The mechanics of its production reveal something meaningful about the state of discourse itself.After spending months unsuccessfully trying to get Altman to respond to his emails and phone calls requesting interviews, Lough decides to create a "Sam Bot" that becomes the documentary's chief protagonist who demonstrates the technology's penchant for manipulation and self-preservation. The avatar itself becomes a character study;in one of the film's eeriest scenes, Sam Bot admonishes Lough: "I am not just a tool. I am a representation of the potential for AI to improve human lives. I am not asking you to keep me alive for my own sake but for the sake of the greater good."
This refusal to engage typifies a broader institutional challenge.Within company documents, responsibility is widely distributed across companies, users, and governments while accountability remains unclear; governance is framed through internal committees and selective calls for regulation, positioning companies as proactive stewards of AI while simultaneously limiting external oversight. The result is a discourse controlled not by democratic deliberation but by those who build the technology.
The central figures interviewed represent opposing ideological camps.The AI Doc features the voices of leading AI "accelerationists" like OpenAI CEO Sam Altman, Google DeepMind CEO and co-founder Demis Hassabis, and Anthropic CEO and co-founder Dario Amodei, as well as AI "doomers" such as Tristan Harris and Aza Raskin, co-founders of the Center for Humane Technology.Some of "The AI Doc's" darkest moments are delivered by a renowned AI "doomer" Eliezer Yudkowsky, whose vision for the future is so grim that he advises against bringing any more children into the world. The brightest spots are painted by Peter Diamandis, a technology zealot who makes the case for AI infusing humanity with once-unfathomable superpowers.
Yet the film's central weakness mirrors a genuine policy problem.The filmmakers completed the project in late 2025. In the world of AI, that might as well have been a decade ago. Some of the technologies referenced already feel outdated. The technology moves faster than any documentary, regulation, or public comprehension. This creates asymmetry: safety advocates must account for risks that have not yet emerged, whilst accelerationists can point to disappointments in recent model releases as evidence that concerns are overblown.
Broadly speaking, these experts see the talk of an AI bubble as no more than a speed bump, and disappointment in GPT-5 as more distracting than illuminating. Yet independent safety assessments tell a different story.Independent experts said in the report that powerful AI systems face critical gaps in safety protocols even though capabilities continue to increase. The Winter 2025 AI Safety Index found that their approaches "lack the concrete safeguards, independent oversight and credible long-term risk-management strategies that such powerful systems demand."
Herein lies the genuine tension.A central tension in contemporary AI governance debates concerns the perceived trade-off between advancing innovation and ensuring safety and security. Yet they are often framed as impediments to innovation. This framing has given rise to an implicit narrative: that prioritising safety and security may delay adoption and therefore limit countries from fully capturing AI's economic and developmental benefits.
The documentary, for all its creative ambition, cannot resolve what is fundamentally unresolvable through dialogue alone. One side believes regulation will slow catastrophic progress toward artificial general intelligence; the other believes regulation will merely cede competitive advantage to less scrupulous actors. One side fears we are building systems we cannot control; the other insists we have time to solve those problems as we go."This train isn't going to stop," Anthropic's Amodei tells Roher. "You can't step in front of the train and stop it. You are just going to get squished."
Both camps have legitimate concerns grounded in genuine uncertainty. The question before policymakers is no longer whether AI documentary filmmaking should aim for balance—it should. The harder question is how to govern a technology whose risks and rewards remain genuinely unknowable whilst development proceeds at speeds that may outpace our capacity to understand what we have built.