A new documentary about artificial intelligence managed to get the most powerful people in tech to sit down and talk on camera. Then it pulled its punches.
The A.I. Dilemma—not to be confused with the 2020 Netflix hit The Social Dilemma from the same creative team—has arrived at a moment when the public is simultaneously fascinated and terrified by what AI companies are building. The film features interviews with CEOs and executives from companies at the center of the AI race, including OpenAI, Google DeepMind, and Anthropic. It’s a rare gathering of the people who are actively shaping the technology that could reshape economies, labor markets, and the nature of human cognition itself. And yet, according to critics, the documentary treats these executives more like misunderstood visionaries than subjects who deserve rigorous interrogation.
As Wired put it in a pointed review, the film “puts CEOs in the hot seat—but goes too easy on them.”
A Familiar Playbook, Recycled for a New Crisis
The documentary comes from the Center for Humane Technology, the organization co-founded by Tristan Harris and Aza Raskin, who previously sounded alarms about social media’s corrosive effects in The Social Dilemma. That earlier film was effective in part because it arrived when the public was hungry for an explanation of why Instagram felt so addictive and why political discourse on Facebook had become so toxic. It offered a clear villain—the attention economy—and a roster of regretful ex-employees willing to testify against their former employers.
The new film attempts to replicate that formula for AI. But the dynamics are different this time. The executives featured aren’t disillusioned former insiders. They’re sitting CEOs and active leaders still very much in charge of the companies building the most powerful AI systems on the planet. That distinction matters enormously.
When a former Facebook product manager confesses that the platform was designed to exploit psychological vulnerabilities, there’s an inherent credibility. The person has nothing left to lose. When a current CEO discusses the risks of their own product while simultaneously raising billions of dollars to scale it, the calculus is different. The audience is watching someone who has a direct financial interest in how the conversation lands.
Wired’s review argues that the filmmakers failed to press hard enough on this tension. The documentary reportedly allows executives to frame AI risk in abstract, almost philosophical terms—existential threats, alignment problems, the potential for superintelligence—without adequately confronting the more immediate and concrete harms already unfolding. Job displacement. Deepfakes flooding elections. AI-generated content polluting the information supply. Copyright lawsuits piling up from artists and authors whose work was ingested without permission to train the very models these CEOs are selling.
These aren’t hypothetical scenarios. They’re happening now.
The film’s approach raises a question that media critics and technologists have been wrestling with for years: Can the people building a potentially dangerous technology be trusted as the primary narrators of its risks? The documentary seems to believe they can—or at least that their participation is valuable enough to justify a softer editorial hand.
There’s a pragmatic argument for this. Getting Sam Altman or Demis Hassabis on camera is genuinely difficult. These executives are lawyered up and media-trained. Securing their participation likely required some implicit or explicit understanding about tone. Documentarians face a classic trade-off: access versus accountability. The film chose access.
But access without accountability produces something closer to a corporate keynote than journalism. And at a moment when the public desperately needs clear-eyed reporting on AI, that’s a significant missed opportunity.
The Risk of the “Reasonable CEO” Frame
One of the more subtle problems with the documentary, according to Wired’s analysis, is the way it positions tech executives as thoughtful stewards who are genuinely worried about the technology they’re building. This framing isn’t necessarily wrong—many of these leaders do express concern publicly and privately about where AI is headed. But it can function as a kind of rhetorical shield.
When a CEO says “I’m terrified of what this could become,” the statement does two things at once. It signals awareness and responsibility. And it preemptively deflects criticism: How can you attack someone who already agrees with you that the risks are real?
This is a pattern that played out during the social media era too. Mark Zuckerberg testified before Congress multiple times, acknowledged problems, promised reforms. The platforms continued largely as before. The “concerned executive” posture became a communications strategy, not a commitment to change.
The AI industry appears to be running a similar playbook. Companies publish safety research. They sign open letters about existential risk. They hire alignment researchers and ethics teams. But they also continue racing to release ever-more-powerful models on compressed timelines, often cutting safety evaluations short in the process. OpenAI’s own safety researchers have publicly resigned over what they described as insufficient attention to safety testing before product launches. Anthropic, which markets itself as the safety-focused AI lab, is still in an arms race for scale and market share.
A documentary that simply lets these executives articulate their concerns—without confronting the gap between their words and their companies’ actions—ends up reinforcing the very narrative the industry wants told. We’re the responsible ones. Trust us.
The timing of the film’s release also matters. AI regulation is being debated in Washington, Brussels, and capitals around the world. The European Union has passed its AI Act. The U.S. is still sorting out its approach, with competing bills and executive orders creating a patchwork of half-measures. In this environment, a widely viewed documentary shapes public perception—and by extension, political will. If the dominant cultural narrative is that AI leaders are already worried and working on the problem, the urgency for external regulation diminishes. That benefits the industry enormously.
Recent reporting underscores just how contested the regulatory terrain has become. AI companies have dramatically increased their lobbying spending in Washington, and several have hired former government officials to lead their policy teams. The argument they’re making to lawmakers echoes the documentary’s tone: the technology is powerful and risky, but the people building it are best positioned to manage those risks. Heavy-handed regulation, they argue, would stifle innovation and hand advantages to China.
It’s a compelling argument. It’s also self-serving. And a documentary that fails to interrogate it is doing the industry’s work for free.
None of this means the film is without value. Getting major AI executives on the record—even in a friendly setting—creates a historical document. Their statements can be measured against future actions. Their predictions can be tested. And for a general audience still trying to understand what large language models actually do and why people are worried, the documentary offers a starting point.
But a starting point shouldn’t be mistaken for a destination. The most important AI stories right now aren’t being told by CEOs in well-lit interview rooms. They’re playing out in courtrooms, where the New York Times and other publishers are suing OpenAI over training data. In labor markets, where companies are quietly replacing workers with AI systems. In schools, where teachers are struggling to distinguish student work from machine-generated text. In elections, where synthetic media is already being used to mislead voters.
These stories don’t have the cinematic appeal of a billionaire CEO contemplating humanity’s future. But they’re where the actual consequences of AI are being felt by actual people. A documentary that centered those stories—while still featuring executive perspectives as counterpoint—would have been braver and more useful.
What the Film Gets Right, Despite Its Blind Spots
To its credit, the documentary does succeed in conveying the sheer speed of AI development. The pace at which capabilities have advanced since the release of ChatGPT in late 2022 has stunned even people inside the industry. Models that couldn’t pass a bar exam two years ago now outperform most human test-takers. Image generation has gone from producing surreal, six-fingered nightmares to photorealistic output that’s increasingly difficult to distinguish from real photographs. And multimodal systems that combine text, image, audio, and video understanding are arriving faster than anyone predicted.
This velocity is genuinely important for the public to understand. It shapes the timeline for every downstream question—about jobs, about safety, about governance. The film apparently communicates this effectively, even if it doesn’t always follow through with the harder questions that the speed raises.
So where does that leave audiences? Informed, but not challenged. Aware of the stakes, but not of the full picture. The documentary is likely to be widely watched and widely discussed, which gives it influence. But influence without rigor is a dangerous thing—especially when the subject is a technology that its own creators describe as potentially the most consequential invention in human history.
The next great AI documentary hasn’t been made yet. When it is, it should include the executives. But it should also include the workers who’ve lost their jobs to automation, the artists whose styles have been replicated without compensation, the researchers who left major labs because they felt safety was being deprioritized, and the regulators trying to write rules for a technology that evolves faster than legislation.
That film would be harder to make. It would probably lose some executive access. But it would be closer to the truth. And right now, the truth about AI is far more complicated—and far more urgent—than any CEO is willing to say on camera.
WebProNews is an iEntry Publication