No One Asked Me.

In March 2026, a Canadian Senator stood up and announced that only 12% of Canadians currently use AI — and that "adoption moves at the speed of trust." She meant it as a call to action. A reason to build a national strategy, spend public money, appoint committees, launch working groups, and consult stakeholders.

I am in that 12%. I use AI every day. It is not a curiosity or a productivity hack for me. It is infrastructure — the analytical partner inside the product I built from scratch, with no institutional support, no government grant, and no invitation to any consultation I know of.

The strategy is called "AI for All." I'd like to take that name seriously. Because right now, it is AI for some: large telecom companies, large banks, large universities, large unions, and the large technology companies whose products the rest of us are supposed to adopt and trust. "All" does not yet include people like me.

My name is Alexandra Kitty. I am an author of 21 books, an educator, a journalist, a researcher, and the founder of KlueIQ — a true crime gaming company based in Hamilton, Ontario, that uses AI as its core investigative engine.

KlueIQ is not a game about AI. It is not a demo of what AI could theoretically do. It is a working product — nonfiction, interactive documentary games built on police and court files, where the AI functions as a constrained evidence-analysis partner for citizen sleuths. Players don't interact with a gimmick. They interact with a tool that forces them to reason, weigh evidence, and confront the ways systems fail victims.

I designed the AI's role deliberately non-magical. It does not declare guilt. It does not name perpetrators. It surfaces contradictions, flags missing information, and asks better questions. This is not because I couldn't build something flashier. It is because I thought carefully about what AI should and should not do when real people's lives and deaths are the subject matter.

That kind of careful, principled, small-scale, high-integrity AI deployment is exactly what every policy document I've read claims to want. It is also exactly what none of them include when they cite adoption figures, write legislation, or fill committee seats.

Statistics Canada reported that 12.2% of Canadian businesses used AI to produce goods or deliver services in Q2 2025. By 2026, that figure had risen to around 16% of businesses — roughly doubling in two years. Those are the headline numbers that show up in Senate speeches and ministerial briefings.

But look at how those numbers are built. "Business AI use" primarily captures large enterprises — firms with the procurement budgets, IT departments, and legal teams needed to integrate AI into formal workflows. It does not meaningfully capture the sole proprietor, the independent studio, the boutique label, the one-person research operation that woke up three years ago and quietly rebuilt their entire creative practice around AI tools.

We are the ones who have the most direct, unmediated relationship with AI — no enterprise licence, no corporate policy, no IT filter between us and the technology. We are also the ones who bear the most risk when that technology fails, when its outputs are wrong, when copyright law is unresolved, when data governance is opaque. And we are the ones who figured it out anyway, because we had to, because our work depended on it.

When Statistics Canada counts me, I am a rounding error. When the Senate discusses AI adoption, I am an abstraction. When "AI for All" allocates $200 billion toward economic growth, the growth they're modeling doesn't include what I'm building. It includes what I might someday buy from someone else — a larger platform, an enterprise tool, an AI-as-a-service product from a company that was in the room when the strategy was written.

I want to say something that may sound harsh, but I believe it's true and I believe policymakers need to hear it from someone who operates at the ground level: the government does not know how AI should be integrated into the arts, into investigative journalism, into independent creative practice, into the true crime genre, or into citizen-facing education. It cannot know, because it has not done it.

This is not a criticism of intelligence or intent. It is a structural observation. Governance of a technology is not the same as expertise in deploying that technology for human and social purposes. The people best placed to advise on how AI should work in investigative storytelling are the people who have built investigative storytelling with AI — not those who have written policy frameworks about why it would be beneficial to do so.

The current approach to AI regulation in Canada trends toward designing rules for a technology that official Ottawa has mostly observed from a distance, based on testimony from parties who have a financial interest in how those rules are written. That is not consultation. It is ratification of existing power.

A more humble government — and I use that word precisely — would acknowledge that it cannot design AI policy for the arts and independent creators without the arts and independent creators in the room. Not as token voices. Not as "public comment" at the end of a process already concluded. At the table, at the beginning, when the questions are still being formed.

Let me be concrete, because policy discussions about AI tend to stay abstract until the abstraction becomes legislation that affects real work.

In KlueIQ, AI does not generate creative content. It does not write the story. It analyzes the evidence structure of a case — the police files, the court records, the timeline of events — and surfaces where the official narrative breaks down. It asks: what's missing here? Whose account contradicts whose? What question was never asked? It is a research engine, not a storyteller. It is a tool for teaching people to think like investigators, not a replacement for investigation.

Building this required making dozens of decisions that no policy document helped me make. What should the AI be allowed to conclude? What should it explicitly refuse to do? How do I prevent players from using it to harass real people connected to open cases? How do I design the interface so that the AI's limitations are as visible as its capabilities — so players learn to distrust it appropriately, not just trust it blindly?

These are not the questions being asked in Ottawa. Ottawa is asking about economic growth, job projections, and national AI sovereignty. Those are real questions. But they are not the only questions. And the people who have already been answering the harder, more granular questions at the craft level — alone, without support, often at personal financial risk — have not been asked to contribute what they know.