AI in public services: which decisions should remain human?

The website’s future scenarios distinguish between AI that advises people and AI that has final authority. A practical way to explore that difference is through a public service.

Imagine a city using AI to allocate a limited budget between flood protection, housing and public transport. AI could compare options and explain trade-offs, but residents may disagree about which priorities matter most.

Which decisions would you delegate to AI, and which would you reserve for elected representatives or citizens? What would make a right to appeal meaningful: an independent reviewer, a deadline for a response, the power to pause a decision, or something else?

There are trade-offs on both sides: too little review could leave harmful decisions unchallenged; too much could delay useful action. What arrangement would earn your trust, and what evidence would make you change your mind?