Relocating far away because of AI

It works if the scenario is an abrupt reduction in sunlight and you have moved to the southern hemisphere, or if it is a pandemic and insularity buys you time. None of that is what people have in mind when they say they are moving because of AI.

It does not work if you cite the paper everyone cites. The first systematic review of which places on the planet are most resilient to global catastrophic risks analyses seven risks —nuclear war, asteroids, eruptions, large-scale cyberattacks, electromagnetic pulse, geomagnetic storms and pandemics— and AI is not among them. Its conclusion, in so many words, is that there is no place on Earth resilient to all of them, and although Australia comes out as the best positioned across the widest range of scenarios, even for Australia international cooperation and trade remain essential ^[580]No Place to Hide? Regional Resilience and Vulnerability to Global Catastrophic RiskJehn, Florian Ulrich; Rössler, Maximilian; Kemp, Luke et al. · 2026 · paperView source ↗verified through CrossrefAccessed on 9 September 2026.

Evidence. Low for what is being asked of it. 80,000 Hours’ own civilisational resilience profile warns that there is very little research on these interventions and that all of it could be wrong ^[5]Civilisation resilienceKoehler, Arden · 2026 · reportView source ↗Accessed on 9 September 2026.

Cost. The highest on the list, in money and in everything else: work, family, language, local network.

What it does NOT solve. None of the mechanisms that motivate the move. The decisions that stop going through you follow you to any country. And there is a documented irony: moving dismantles the neighbour network that does predict who recovers from a disaster, which is the only thing with hard evidence in this whole block.