As declining geopolitics affects technology policymaking globally, the risk for startups expanding beyond national borders is rising fast too. While national laws might be able to deal with simpler AI applications and LLMs, frontier models may require global rules - and an international body to oversee them. The ICAO (Aviation), CERN (Nuclear), IAEA (Atomic Energy), are organisations from the past that embody ways to govern a powerful new technology. And these have mostly worked.

But getting a global agency to oversee AI anytime soon might be quite the tall order.

So what can national policymakers do to create a framework that allows for startups at home to not only attract foreign capital, but expand beyond their borders easily, and at the same time contain risks from imported AI systems causing significant harm?