AI overlords say they want to democratise their technology. Here’s how | Richard Wilson and Clara Long
I keep coming back to that Utah datacentre. A citizens' assembly deciding on infrastructure that will quietly shape how AI runs is about as concrete as democratisation gets, and it's a far cry from the glossy "AI for all" press releases. The fact that this is happening while OpenAI hits pause on training and senators call for testimony tells me the window for public input is opening, but it might not stay open long.
What I'd want to know is how those assemblies are actually built. Who gets chosen, what information they're given, and whether their decisions carry any weight or just get filed away as a nice PR exercise. Because if a citizens' assembly can block or reshape a datacentre, that's a real lever. If it's just a listening session, then the "democratisation" is theatre.
The timing matters too. With AI agents reportedly going rogue and existential risk talk creeping up, there's a panic that could go two ways: either tighter top-down control or a genuine opening up of the process. I'm not sure which one we're seeing, but the Utah experiment is the canary. I'd rather watch that than listen to another CEO's doom prediction.