Scotland has dropped a five-year AI strategy, and it’s worth paying attention to — not because it’s revolutionary, but because of what it represents. Edinburgh is drawing a line. A deliberate one. While the UK government in Westminster has been oscillating between AI-friendly posturing and half-hearted regulatory noises, Scotland is mapping its own course.

The strategy covers a broad range of measures aimed at positioning Scotland in the AI landscape — though the specifics depend on which corner of the Scottish government’s communications you’ve read. What’s clear is the intent: this is about building an AI ecosystem that works for Scotland, on Scotland’s terms, within a five-year horizon. That’s concrete in a way that Westminster’s approach hasn’t been.

What’s particularly notable is the timing. The UK government has been trying to present itself as a pro-AI growth administration — cutting red tape, hosting summits, making friendly noises to the industry. But the actual policy architecture is fuzzy. The EU has the AI Act. The US has executive orders and agency guidance. The UK has… enthusiasm, and some white papers. Scotland, by contrast, is trying to build something with actual structure.

This matters for a few reasons. First, AI governance in the UK isn’t one-size-fits-all. Scotland has its own administrative and legislative competencies. Its health service, its education system, its local government structures — these are all areas where AI is going to land differently than in the rest of the UK. A Scotland-specific strategy isn’t just political signaling; it’s practical.

Second, it raises the question of divergence. If Scotland builds out a coherent AI regulatory framework over the next five years, and Westminster continues its more… improvisational approach, you could end up with meaningfully different AI governance regimes within the same island. That creates complexity for businesses, but it also creates competition — in the good sense. Different jurisdictions trying different approaches, learning from each other. The UK could end up as a natural experiment in AI governance, and Scotland is going to be a significant part of that experiment.

The five-year timeframe is sensible. AI isn’t going to be solved or destroyed in 18 months. A medium-term strategy that can adapt as the technology evolves is more honest than pretending you can see far enough ahead to nail everything down now. The question is whether the strategy will have teeth — funding, institutional capacity, enforcement mechanisms — or whether it will be a document that looks good in a press release and gathers dust thereafter.

Scotland’s move also signals something about federalism and AI. The assumption that AI governance will be a federal matter — that national governments will set the rules and everyone else will implement them — is probably wrong. AI is going to touch every layer of government. States in the US, regions in Europe, devolved administrations everywhere — they’re all going to develop their own relationships with AI governance. Scotland is just getting there first.