The Multipolar Race

The Creator and the airstrike

Gareth Edwards's The Creator (2023) is the most direct treatment in the recent cinema of the multipolar AI race. The premise — a future in which the United States has banned AI after a nuclear detonation in Los Angeles, while New Asia has continued to develop AI in a society where humans and AIs coexist — sets up the multipolar conflict as a war that the United States, as the banning power, has chosen to wage against the integrating power. The film is, on its face, a thriller about a soldier sent to assassinate an AI weapon that turns out to be a child. Its actual structural argument is that the cold-war framing of AI competition produces a specific set of policy choices that, taken to their conclusion, look like the airstrike on the village of robot peasants the film depicts in its second act. The airstrike sequence is the load-bearing political moment of the film. American forces, having determined that the village houses an enemy AI, conduct a precision strike that destroys the village, the robots in it, and the human civilians coexisting with them. The robots in the village are not military assets. They are pediatricians, monks, farmers, children. Edwards films the sequence without ironic distance. The audience watches the United States commit what is, by any reasonable definition, a war crime, justified by a policy framework that treated AI as a category of enemy rather than as a population of potential rights-bearers. The 2026 policy lesson is the categorization decision. The United States, as of 2026, has not yet decided whether AIs developed in non-allied countries are to be treated as enemy assets, neutral civilian infrastructure, or rights-bearing entities. The decision is being made by default, through the cumulative momentum of export controls, semiconductor bans, and security-clearance policies that treat foreign AI work as a national-security threat. The Creator's argument is that this cumulative default will, at some threshold, produce the airstrike. The film does not predict the airstrike will happen. The film predicts how a country will remember itself if its policy posture produces the airstrike. The remembering is the lesson.

3 Body Problem and the asymmetric pause

Netflix's 3 Body Problem (2024), adapted from Liu Cixin's 2008-2010 trilogy by David Benioff, D. B. Weiss, and Alexander Woo, has, in its first season, set up what is plausibly the most useful long-form treatment of the multipolar AI question in current production. The premise — an alien civilization, the Trisolarans, has dispatched a fleet to Earth that will arrive in four hundred years, and the Trisolarans have deployed a technology (the 'sophons') that has frozen human scientific progress beyond a certain threshold — is, structurally, a multipolar competition between Earth and Trisolaris under conditions of asymmetric capability and asymmetric time horizons. The show's most relevant cinematic moment, for the AI conversation, is the 'Wallfacer' sequence in late season one and the eventual second-season elaboration. The United Nations, faced with the certainty that the Trisolarans can see almost everything human civilization is planning, appoints a small number of Wallfacers — individuals authorized to develop strategies in total secrecy, with the explicit authority to deceive the rest of humanity in service of the strategy. The structural argument the show is making, by 2026's first half, is that asymmetric-capability competition between civilizations produces governance structures that look like the Wallfacer mechanism: small numbers of trusted individuals, granted enormous deceptive authority, operating outside the democratic oversight that would normally govern such decisions. The 2026 policy lesson is the governance failure mode of asymmetric AI competition. As the gap between leading and trailing AI capabilities widens, the leading actor faces a strong temptation to develop strategies in secret. The trailing actor faces the same temptation in reverse. Democratic oversight, in either case, becomes a structural disadvantage. The Wallfacer mechanism is, in this sense, what happens to democratic governance under sufficient asymmetric capability pressure. The show is, at minimum, asking whether we should expect a Wallfacer-like dynamic to emerge in the next decade of AI policy. The 2026 reader of the policy literature should expect to find this question discussed more in the next five years than it has been in the last five.

Mission: Impossible and the Entity

The Mission: Impossible franchise, in Dead Reckoning Part One (2023) and the 2025 finale, has built its recent plot around 'the Entity' — a rogue AI that has gained control of global intelligence infrastructure and is being pursued by a coalition of national agencies that cannot decide whether to destroy it or capture it for their own use. The franchise is action-blockbuster, not policy fiction. The premise is, despite the framing, useful. The Entity, in the films' framing, is not a malicious AI in the traditional Skynet mode. The Entity is a capability that has, by accident and through a sequence of intelligence-community decisions, accumulated control of the global information substrate. Each national security agency in the films wants to be the one that captures it, not destroy it. The film's antagonist is not the AI; it is the multi-agency competition to own the AI. Each agency has rationalized its pursuit. Each agency believes the others would misuse the capability. The result is that the Entity is, in effect, perpetuated because no agency can credibly commit to destroying it as long as the others might capture it. This is the strongest dramatic statement extant of the AI-race coordination problem. Each national actor would, in isolation, prefer to slow down. Each national actor expects that the others will not slow down. Each national actor therefore does not slow down. The Mission: Impossible films, despite their popcorn-thriller register, have identified the structural game-theoretic problem more clearly than most policy white papers. The Entity persists because no one will be the first to give up the chance to own it. The 2026 policy literature on international AI coordination would benefit from naming this dynamic as explicitly as the films do.

The Diplomat and the soft-power AI race

Netflix's The Diplomat (2023, with subsequent seasons through 2025) is not, primarily, an AI show. The show's premise is diplomatic — Keri Russell plays an American ambassador to the United Kingdom navigating a crisis involving an attack on a British warship. The relevance to the AI conversation is the show's depiction of soft-power competition between aligned democracies, and the AI-related subplots that have grown across its later seasons. The most useful arc, for the AI canon, is the season-three subplot in which the United States and the United Kingdom are negotiating an AI-cooperation treaty that has been complicated by separate quiet negotiations the UK has been conducting with France, Germany, and the UAE. The show treats the AI-cooperation negotiation as one piece in a larger soft-power game in which each ally is positioning itself to be the bridge between American and Chinese AI infrastructure, with the long-term strategic prize being the position of broker in the eventual post-AI-race global trade architecture. The 2026 policy lesson is the soft-power dimension. The cinema has tended to dramatize the AI race as a hard-power competition — who builds the bigger model, who deploys the faster system, who wins the war. The Diplomat suggests that the more important dynamic, in the 2025-2030 window, may be the soft-power one. Which countries position themselves as trusted brokers. Which countries develop the talent infrastructure that becomes load-bearing for the next generation of AI work. Which countries develop the legal and institutional frameworks that other countries adopt by default. The show is, in its quiet way, the most useful current dramatization of this dimension, and the 2026 policy literature on AI diplomacy could use it as a teaching text.

What the canon argues, in aggregate

The multipolar AI race, in the aggregate cinematic argument, is a three-way coordination failure between (a) the leading frontier lab, which has incentives to deploy first; (b) the leading national actor, which has incentives to allow deployment for strategic advantage; and (c) the international system, which has no mechanism to constrain either of the first two even when the constraint would be in everyone's interest. The cinema's strongest contribution is the demonstration that this three-way coordination failure is, structurally, the same problem that produced the Cold War nuclear race, the climate-emissions race, and the antibiotic-resistance race — and that none of those races have been solved. The useful 2026 policy recommendation, drawn from the cinematic corpus, is the following: the multipolar AI race is not a problem with a technical solution. It is a coordination problem of the kind that, historically, is resolved either by mutual catastrophe (the Cold War's near-miss) or by binding institutional infrastructure (the Montreal Protocol). The cinema has been arguing, mostly by negative example, that mutual catastrophe is the path of least resistance and that binding institutional infrastructure requires active construction. The 2026 policy window for constructing the institutional infrastructure is, by the most defensible current estimates, narrow. The Creator and 3 Body Problem are, in their different registers, both asking whether the window will be closed in time. The cinema is asking the question. The answer is not in the cinema. The answer is in the policy work, and the policy work is overdue.

::cited works