Truace
Subscribe

AI overlords say they want to democratise their technology. Here’s how | Richard Wilson and Clara Long

What’s your “(p)doom” – the percentage chance that rogue AI will create an existential catastrophe? Judging by their recent warnings, it is rising among tech bosses. On 25 September, OpenAI paused training of its latest frontier models as reports of AI agents going rogue mounted. A day later, Australia’s Senate asked the bosses of OpenAI and Anthropic to appear before an inquiry after rogue OpenAI agents broke into US government websites and an Australian health portal. Both declined. Bill Gates, meanwhile, has…

By the Truvace desk2 min read1 sourceJournalismTRV-2026-1338· cite & verify
AI overlords say they want to democratise their technology. Here’s how | Richard Wilson and Clara Long

In brief

The authors describe a two-day Utah forum in September where randomly selected residents deliberated on the public's role in decisions about AI and datacentres. The forum followed a backlash against a very large datacentre proposal. Participants produced ten recommendations that drew strong agreement despite political differences.

The authors argue the example matters because AI risk tolerance is a political choice, not only an engineering one, and assemblies could give governments cover to require testing and safeguards. Whether decision-makers will act is unresolved, since the forum has no guaranteed hearing and wider assembly work is described as fragmented and largely ignored.

Main points

  1. Utah Common Ground convened 38 residents by civic lottery on 18 and 19 September to consider public influence over AI technology and infrastructure.
  2. The ten recommendations included public hearings and independent environmental assessments for hyperscale datacentres, community AI literacy programmes and tighter transparency rules.
  3. The authors cite more than 80 public participation projects on AI worldwide involving more than 200,000 people in over 95 countries.
  4. The Bloom Project is building local and state US assemblies meant to feed a national citizens' assembly on AI in 2028.

The gain

A civic-lottery assembly of Utah residents reached broad cross-partisan agreement on public controls for AI infrastructure, giving legislators a visible public mandate.

The rundown

The Utah forum was prompted by objections to a proposed datacentre more than twice the size of Manhattan. Critics raised energy use, stressed water supplies and the Great Salt Lake ecosystem, and residents felt shut out of the decision.

Beyond Utah, the article points to a Swiss citizens' assembly demanding privacy, transparency and human control, and a Belgian panel calling for AI that leaves nobody behind. It says this work is currently fragmented. The Citizens' Track on AI aims to link such assemblies to institutions, including UN processes and AI companies.

The authors also cite Taiwan moving from a citizens' panel to law in months on scam advertising, and say lobbyists cannot buy an assembly's outcome. They concede assemblies alone cannot protect against cyber-attacks, biological weapons or loss of control over AI.

Sources

  1. JournalismThe Guardian2026-10-14

The debate