“La P8stée” (2027) starring Bobcat and P8stie

AI catastrophic risk communication today is dominated by research papers, policy briefs and short explainer videos. There's still a wide audience that doesn't know about it. We are trying to reach that audience through the power of images and impress the idea of x-risk into the minds of people in the same way that Threads, The Day After and The War Game scared people (and policymakers) into thinking about how to prevent nuclear war.

Nuclear risk had its films. x-risk from AI does not. What exists is either blockbuster-style distortion (robopocalypse tropes that make the real threat models easier to dismiss) or educational material for the already-converted. I have worked inside AI companies on trust and safety. This includes foreign interference by state actors, which is the exact layer the AI 2027 scenario runs on. The film stays faithful to those threat models, not movie logic.

Within the x-risk portfolio, narrative cinema is an essentially unfunded channel, and the minimum amount of funding ($50,000) ships a complete film. We expect three audiences and three effects:

• Festival programmers and art-cinema viewers. To people not convinced of any x-risk or p(doom), x-risk registers as hype from the people selling AI products. Festival selection is legitimizing: programmers, critics, and other filmmakers treat it as a serious subject rather than a sales pitch, and their audiences follow. These viewers are also disproportionately the people who make more culture (writers, academics, journalists, curators), and so the idea propagates.