Beyond Borders #3: The Year the World Started Arguing About Who Controls AI
I wanted to write about AI this week, and the more I read, the more I realised the actual story isn't "AI is getting smarter." It's "the world just started fighting over who gets to decide the rules for it" — and that fight is now a bigger diplomatic story than most actual treaties being negotiated this year.
Here's the headline development: 2026 is the year AI governance went genuinely global for the first time.
The UN backed something called the Global Dialogue on AI Governance and an Independent International Scientific Panel on AI — the first real forum where nearly every state gets a seat to debate AI's risks and rules.
That sounds like good news, and in one sense it is. But the three biggest players walked into that forum wanting completely different things. The EU wants a rights-based, risk-based regulatory model.
The US favours voluntary standards, prioritising innovation and security flexibility. China promotes "inclusive cooperation" while keeping tight state control over data and deployment. Same forum, three incompatible starting points.
The Battle of the "AI Stacks"
Analysts keep using this phrase — "AI stacks" — to describe something that's easy to miss if you're only thinking about chatbots: the actual physical and regulatory infrastructure underneath AI.
Compute power, cloud storage, microchips, the rules governing all of it. The US, EU, and China are each trying to get other countries to build on top of their stack rather than a rival's.
The White House has explicitly made it policy to export the American AI stack to other countries, in part through funding support.
Whichever stack a country ends up standardised on will shape its economy, its security relationships, and its politics for a long time — this is genuinely closer to picking an alliance than picking a vendor.
This is also why the "sovereign AI" movement is exploding right now. Countries are pouring huge amounts of capital into their own national AI infrastructure, on the theory that depending entirely on a foreign AI provider is now a real economic and security vulnerability — not a hypothetical one. The logic is blunt: control AI, or be controlled by whoever controls it for you.
What This Actually Does to People's Lives
Underneath the governance fight is a much more immediate economic one.
The OECD estimates AI could add somewhere between 0.4 and 1.3 percentage points to annual productivity growth in high-exposure economies over the next decade — which sounds small until you realise that's the kind of number that reshapes entire labour markets over time.
But the gains aren't landing evenly.
Analysis of who's actually using advanced AI tools shows high-income economies are overrepresented relative to their share of the world's working-age population, which means AI could just as easily widen the gap between rich and poor countries as close it.
There's also a genuinely uncomfortable structural question sitting underneath all of this that almost nobody in my generation is being asked to think about: most government tax systems, including the US's, rely overwhelmingly on taxing labour income.
If AI starts doing a meaningful share of the work that used to generate wages, and wages are what governments tax to fund everything else, that's not just a jobs problem. It's a "how does a government pay for anything" problem.
I don't think anyone — not the analysts, not the governments, not the companies — actually knows how this resolves. What I do think is that most of the AI conversations I hear people my age having are about whether it'll write their essays for them. The conversation actually shaping the next decade is about who owns the infrastructure, whose rules apply, and who gets left out of the productivity gains entirely.
Questions to Think About
If a country builds its economy on the "American AI stack" versus the "Chinese AI stack," how much does that choice actually function like a foreign policy alliance rather than a technology decision?
Is a global AI governance forum meaningful if the three biggest powers in the room want fundamentally incompatible rules — or is showing up itself the point?
If AI adoption is currently concentrated in high-income economies, what would it actually take for a country like India to make sure the productivity gains don't just widen the gap it's already trying to close?
Where This Fits Into Beyond Borders
This continues the thread from last week's post on business and supply chains — AI infrastructure is turning into exactly the kind of strategic asset that countries fight over the way they used to fight over oil or trade routes. It's also the clearest example yet of something this series keeps returning to: the technical and the geopolitical are the same story now, told from two different angles.