How should cities govern AI?

What's emerging

Across the conversation, strong consensus emerges around transparency and responsible procurement: residents broadly agree that the city should publish a public register of AI systems, pilot tools in time-limited sandboxes before purchase, and require open-standards clauses to avoid vendor lock-in. There is also wide support for giving residents the right to human review of automated decisions. The genuine divides centre on institutional oversight structures — specifically whether an independent board should have power to pause deployments and whether a dedicated AI ethics office should oversee all departments, both of which split opinion sharply.

• TransparencyStrong cross-group agreement that the city must publicly disclose every AI system it deploys, making its use of automated tools visible to all residents.
• Resident RightsBroad support for guaranteeing that anyone affected by an automated decision can request a human review, protecting individual recourse.
• Procurement SafeguardsConsensus that AI tools should be piloted in sandboxes before purchase and that contracts must include open-standards clauses to prevent vendor lock-in.
• Oversight StructuresSharp division over whether an independent board should be able to pause AI deployments and whether a dedicated ethics office should oversee all city departments.