The title gives the game away pleasantly: this is a documentary that knows it is caught between two irreconcilable positions and has chosen to live in the hyphen between them rather than pick a side. That intellectual honesty is its greatest strength and, ultimately, also the source of its limitations.
The three-act structure is clean and deliberate. The pessimists go first: AI risk analysts and commentators who paint extinction-level futures with genuine conviction. Notably, most of them work at a distance from the actual construction of these systems, which the film does not editorialize about but lets the audience register on their own. Then come the optimists, drawn almost entirely from inside the major AI labs, whose proximity to the technology produces a different kind of confidence, not necessarily more warranted but certainly more specific. The final section lands in the middle ground, asking what responsible development actually looks like in practice rather than in theory.
The most genuinely unsettling insight the film offers is the arms race framing. These companies are not simply competing commercially; they are participating in a global contest for technological dominance in which the incentive to move fast consistently outweighs the incentive to move carefully. The adversarial dynamic between governments accelerates everything and disciplines nothing. Nobody decided to start this race. Nobody has the authority to pause it. The CEOs invited to speak are admirably honest about this: none of them can promise safety, only better protocols and the hope of international cooperation, which is either reassuring or deeply alarming depending on your starting disposition.
The recurring question posed to each interviewee, “do you want children,” is the documentary’s smartest structural choice. It anchors an abstract debate about systems and timelines to something irreducibly personal. Whether or not AI destroys us matters differently when you are thinking about a specific child who will inherit whatever world we build, or fail to build. It reframes the question from civilisational to intimate, and the answers reveal more about each speaker than any policy position they articulate.
Where the film falls short is depth. If you have followed the AI discourse at all, read the think pieces, watched the congressional hearings, or engaged with the public debate, The AI Doc will offer one or two genuinely new perspectives but largely confirm the landscape you already recognise. It is an excellent entry point and a competent survey. It is less useful as a document that pushes the conversation meaningfully forward.
Decent, well-structured, and worth your time if the question matters to you. Which, given the subject, it probably should.