Toronto's Questions — an open research project for a city deciding its future
Toronto votes on October 26, 2026, and the four budget years that follow will lock in decisions — on housing, climate, transit, policing, technology, and money — that the city will live inside for decades. This project exists so that when Torontonians sit down to decide, together, what they want their city to become, nobody can say we're not ready or the information isn't there to be discussed. Over the next few weeks we are building the truth layer for that conversation: the problems named honestly, the right questions asked about each one, and the evidence — with receipts — for anyone to check.
1 · We start from problems, not topics
Most civic research is organized the way institutions are organized: by department, by file, by policy area. People don't live in policy areas. They live in a rent cheque that got bigger, a grocery bill that got heavier, a parent on a wait list, a commute that got longer, a basement that flooded. So the project starts from the problems as they are lived, ranked in the open by scale, severity, trend, equity, leverage, and urgency:
Cost of living. Housing. Homelessness. Climate — adapting to it and cutting our share of it. Transit and congestion. The trade war. Policing and surveillance. The city's own data and AI. The city's broken finances. And behind those nine, every other problem the city carries — from primary care to food insecurity to youth opportunity to the state of the pipes — each with its own page, none silently dropped. The ranking itself is published and arguable; disagreeing with it is part of the conversation. Start anywhere: the problems.
2 · Every problem gets the same honest treatment
Each problem carries four things, in plain language:
The problem, with receipts. What it is, who it hits hardest, how big, which way it's trending — every figure traceable to a primary source. Where our own sources contradict each other, we say so on the page instead of quietly picking a number.
The questions. The most important part of research is asking the right questions. Ours are open and exploratory, and they follow a consistent discipline: Where are we actually at — measured, not imagined? How did it get this way — root causes traced all the way down, including the decisions someone actually made? Who profits from the problem staying unsolved, and what breaks if it's solved? Who does this best in the world — and what do we shamelessly copy? What has already been recommended, by whom, and what happened to it? Who are the best minds and best organizations on this anywhere — including the world-class ones already in Toronto being ignored? And where is the strongest counterargument to our own framing right — what would change our mind?
What already exists. The evidence shelf: a full neutral backgrounder on each issue, a Toronto-specific brief, actionable idea cards, and the deeper research where it exists — all linked from the problem page.
What's unexplored. The gaps, named plainly. Where our shelf is thin, we say it's thin.
3 · How we'll pursue the questions — AI as research staff, humans as judges
This project is built AI-forward, and honest about what that means.
What the AI does. Frontier AI lets a small civic team do what once took an institute: sweep legislation, budgets, audit reports, datasets, academic and grey literature — across jurisdictions and languages — for every question on every problem; collect all existing information and recommendations before deriving anything new; draft syntheses; and keep the whole estate current as the facts move. Volume is not the achievement. Traceability is.
The receipt rule. Every factual claim we publish is logged in a claim claims register with its primary source. Claims are independently re-verified against the primary — not the news story about the primary — before they carry weight, and anything not yet verified is visibly marked as such. When we find our own errors, we correct them in public and keep the record of the correction. A hostile fact-checker should be able to walk from any sentence to its source.
What the AI is not allowed to do. It does not invent facts, estimate figures that should be looked up, or launder advocacy into evidence — positions in a debate are catalogued as positions, never as findings. And the judgment calls — how to frame a problem, which values tension a question should surface, what is fair to say about a named institution — stay human.
The cycle we'll run for every problem, in rank order — starting now, with the cost of living: frame it honestly → collect everything already known and recommended → scan the world for who does it best → synthesize with receipts → publish → invite correction and challenge → update. Research here is a loop, not a report, and the loop starts turning this month.
4 · How we'll find the people who can help Toronto answer these questions
No city answers questions like these alone, and the answers mostly already exist — in other cities' results, in researchers' work, in the practice knowledge of people delivering services here today. So alongside the evidence work, over the coming weeks, we will run a deliberate, systematic world scan, per problem:
Who has actually done it. The cities, countries, and programs that measurably improved this problem anywhere — what exactly they did, what it cost, what failed — named as case studies to learn from and, where they fit, to copy without embarrassment.
Who knows it best. The researchers, institutes, and practitioners whose work on this problem leads the field — anywhere in the world, in any language.
Who is already here. The Toronto and Ontario organizations, frontline experts, and community leaders who hold the ground truth on each problem — including the ones whose work is world-class and chronically ignored.
Who holds the data. The agencies, offices, and datasets a real answer depends on, and what it takes to open them.
The scan's output will be public: an open, growing roster for each problem — organizations, experts, case studies, example projects — built by AI-assisted search, kept honest by human review, and growing week over week. Then we'll use it: named experts and organizations will be invited to review our work, redline it, contribute answers, and challenge our framings before major publications; local practitioners will be treated as the ground truth our syntheses must survive contact with; and candidates for office will be asked to engage with the same questions on the record. If you're reading this and you're one of these people — the invitation is standing.
5 · What we'll publish — rolling, starting with the top of the list
For every issue: a full, neutral backgrounder; a Toronto brief; concrete idea cards a citizen or candidate can act on; and discussion guides for real rooms. For every problem: the living question page, updating as its research wave completes. The next few weeks, concretely: the world scans and recommendation inventories run first on the top-ranked problems — cost of living, housing, homelessness, climate, transit — with expert and organization rosters drafted and invitations out as each one lands; syntheses update with receipts as the evidence comes in; publication rolls problem by problem down the ranked list. All of it free, open, and correctable — built to be forked by any other city that wants to run the same play.
6 · What this is, and isn't
This is neutral civic infrastructure: no party, no candidate, no platform planks. The research doesn't tell Toronto what to decide; it makes the deciding honest. The conversation — thousands of small ones becoming one big one — is where the decisions belong, and the people of this city are the ones who get to make them.