uniteTOlove.ca

The Toronto Robot Fair — Explained Simply

v0.1, 2026-07-19 Date: 2026-07-19 ·

Status: v0.1, 2026-07-19 · RENDER — derived from this library's internal records and this library's internal records (both 2026-07-18), plus the program charter, this library's internal records (2026-07-18). DRAFT, pre-ratification. The charter is written, but the operator hasn't yet answered the three questions this plan needs answered before it can move forward. No real fair has ever been held. This page changes none of that — it only explains it.

This is written for anyone who wants to know what Toronto's plan for talking about robots and automation actually is, without digging through a stack of research files to find out. It explains the plan, why it matters, how anyone would know if it's working, and where it could go wrong — based only on research already done for this program. Nothing below is a promise.

What it is

Robots and AI are already changing what work looks like. The plan starts from a simple belief: nobody has really asked people what they think about that yet. This plan, called AUTOMATA, is not a push to bring more robots into Toronto. It is a plan to help the city talk about robots honestly, see them up close, and decide together what it wants. The goal is to make that choice before it gets made for the city by whoever sells the machines.

The plan has one hard rule under everything else. Automation makes people afraid, and this program is built, on purpose, to never trigger that fear. It does that with three connected parts, and none of them work alone.

First, the Toronto Robot Fair. This is a monthly, local event where people can walk up to real machines. Visitors can touch the ones that are safe to touch. A worker who actually runs the machine explains it — never a sales pitch. Once a year, a bigger world-showcase version caps off the year.

Second, the Literacy Engine. It's short, honest explainers, published wherever the movement already reaches people. It's also small kits, called purok demo kits, that a trained neighbor can carry into someone's living room. And it includes a try at getting free or cheaper access to major AI tools, so people can actually use them, not just read about them.

Third, the People's Deliberation Layer, also called the Living List. It's a growing public list of what Toronto residents say they do NOT want a machine to ever do. It grows from cards people write at the Fair.

At the door of every Fair stands a refused-applications wall. This is a board naming what the movement has already ruled out, no matter how the technology improves. That list includes weapons, spying robots, care that pretends to be a real relationship, and gig-work deals that dress up job losses as innovation. Blank cards sit next to the wall so visitors can suggest more.

One big question the plan leaves open on purpose: who would actually own the robots, if Toronto ever had many of them. That decision is left for a related project to work out. This program's job is only to study the options and show people what exists. The plan is also clear about one thing: none of this works without a real safety net for people who lose work to automation. Exactly what that net looks like is left to that same related project.

The plan also protects work people actually want to keep doing by hand. If an elder wants to keep shoveling their own snow or growing their own food, that stays their choice — a machine doesn't get to decide it knows better.

Why it matters

Start with how much of today's work AI can already touch. The International Monetary Fund puts global exposure at around 40% of jobs. In a wealthy economy like Canada's, that figure is around 60%. This means AI can do some real part of the job, not the whole thing. Statistics Canada's own estimate lands in the same place: about 60% of Canadian workers hold jobs AI touches in some way.

But "AI touches part of the job" is a very different claim from "AI can do the whole job today." Mixing the two up is the single most common mistake in this field. When researchers try to figure out how many jobs AI could replace outright right now, the honest number is much smaller. It's roughly one job in ten to one in four — not four in ten.

Actual, measured job losses so far are smaller again. The clearest evidence comes from freelance work sites: exposed freelancers saw about 2% fewer monthly contracts and about 5% lower earnings after a major chatbot's public release. That's real, but it's narrow. It's mostly limited to gig work, not the whole economy. Nearly three years after that release, one ongoing US tracking project still finds no clear sign that AI has shifted the overall mix of jobs people hold.

Meanwhile, worry is already way ahead of actual contact with the technology. In a survey of American workers published in February 2025, 52% said they were worried about AI's effect on their future at work. But only 16% said they currently use AI in their own job. That gap — between how scared people already are, and how little most of them have actually touched what scares them — is exactly the condition this whole plan is built to answer.

There's a deeper reason the plan exists at all. If the people who will live with these machines don't help design how they get used, someone trying to make money off them will design it instead. And that goes beyond robots — it's also true of banking, surveillance, and decision-making systems being built the same way right now.

How would we know it's working

This plan is a research-and-conversation program, not a construction project. So "working" means something different than a building going up. It has its own staged plan, and its own honest rules for when to stop.

The stages. Nothing launches all at once. First, the operator has to answer three open questions before the plan can formally proceed. Next comes a pass to double-check every uncertain fact this research still flags. Then come the actual research write-ups. After that, the literacy materials and fair playbook get checked against safety and accessibility rules. Only then does a trial fair get planned on paper. A real, in-person pilot fair happens only after that paper plan holds up, and only once safety and accessibility reviewers sign off.

The smallest real test, if it happens. One Saturday, at a library branch or community center — the exact venue hasn't been picked yet. Five stations make up the day: a small, charming machine; one genuinely useful machine, explained by the person who actually uses it; and a hands-on table for kids. Two more round it out: a station called The Sidewalk Test, honestly telling the story of Toronto's robot-sidewalk history, and the refused-applications wall. Visitors leave one card at the door, answering a single question: "What do you never want a machine to do?" That card box becomes the seed of the public Living List.

What would count as failure, decided in advance. These stopping points are fixed before anything launches, not decided afterward.

The Fair stops or scales back if:

and a second injury from the same type ends it for good.

event. The Fair pauses until it is.

format is declared a failure and rebuilt from scratch.

and if it happens across several companies, only movement- and worker-run machines stay.

back into a smaller, borrowed-space version instead of limping along.

The Literacy Engine stops or resets if:

The editor is replaced.

walks away, says why in public, and uses open, free tools instead.

checked. That kit is pulled back and rebuilt before any more sessions run.

The Living List stops or gets rebuilt if:

until there's an actual way that input leads somewhere.

is pulled back publicly and the process rebuilt.

rebuilt, and from then on, cards are always published exactly as written.

What could go wrong

The Fair could quietly become free advertising. The plan owns almost no robots of its own. A real fair would have to fill its floor with other companies' machines. That risks turning an honest-learning event into a marketing channel, no matter what the signs say. The plan's answer is the refused wall, workers instead of salespeople doing the explaining, and the rule that bans any company caught collecting data or pushing sales. Whether that actually holds up has never been tested, because no real fair has happened yet.

A number could get oversold. The evidence behind this program includes a fact-check of a specific 40%-of-jobs claim. The honest verdict is mixed. It's fair if it means AI touches part of the job. It's overclaimed if it means AI can already do the whole job. If anyone repeats the bigger, scarier version of that number without the qualifier, it breaks this program's own honesty rule.

One founding story hasn't been fully confirmed yet. Part of the Fair's design leans on one claim: that Toronto's council banned delivery robots from sidewalks after accessibility advocates objected. That detail still needs a final check — the exact rule, the date, and whether it's still in effect — before anyone treats it as settled fact. The Sidewalk Test station would retell that story honestly either way, and the same accessibility community would help design the whole event, not just review it after the fact.

A robot at the Fair could still become a privacy risk. The idea behind the plan's rule is simple: a robot is a camera on wheels. The design promises that recording stays off by default, that nothing leaves the room, and that face-scanning is never used. But that promise only holds if it's actually enforced at every single event. That hasn't been tested in practice yet.

Robots meant to help elders or people needing care are the most contested ground here. Some care robots show real, modest benefits — better mood, less agitation. But the studies behind that finding tend to be small, and not always well reported. Researchers also disagree, in a live argument nobody has settled, about whether such robots can feel fake or make people feel less like adults. But just as many people who actually live with these robots say they don't feel that way, and would rather keep the choice than have it taken away. That disagreement matters most for people whose ability to say yes or no may not be steady — and researchers admit they don't have a clean answer for that yet either. This program's own rule — never sell a robot as a substitute for a real relationship — is stricter than what the evidence currently requires. But the underlying disagreement itself isn't settled by anyone yet, this program included.

The humanoid-robot side of the industry is mostly hype right now, by its own reported numbers. Money flowing into humanoid and robotics companies roughly tripled between 2023 and 2025. But there's still almost no public revenue to show for it anywhere in that part of the sector. That gap comes from company statements and press reports, not numbers checked by anyone outside the industry — this program's own evidence review flags that limit plainly.

Nothing above has actually launched. The charter's three questions to the operator are still open. No real fair has been scheduled. None of the safety or accessibility reviews the plan requires have happened yet.

Receipts

Everything above comes from three documents, all dated 2026-07-18 and still marked draft:

and the three questions still waiting on the operator.

Fair, the Literacy Engine, and the Deliberation Layer, including the models considered and rejected, and every place the research itself flags a fact as still needing a final check.

what's actually been measured about job losses so far, and the honest state of the research on humanoid robots and care robots.

Every number and claim in this explainer traces back to one of those three files. If a figure above doesn't have a citable source in those files, it isn't in this explainer either.

There's no fair to attend yet, and no list to sign — this plan hasn't been ratified, and nothing above has happened in the real world. The one honest action available right now is to watch for two things: whether the charter gets ratified, and whether an actual Saturday fair ever gets scheduled. If one does, the real test of this page is simple — show up, put your hand on a machine, and write your own card for the list of what you never want a machine to do.