I Got Thrown Out of the Library of Parliament
On the commons, radical availability, and the conversation Canada isn't having about its own information.
There’s a version of this story that sounds dramatic.
I got thrown out of the Library of Parliament.
That’s not exactly what happened.
But it’s not wrong either.
The Thing I Was Chasing
I wanted access to data.
Not classified data. Not sensitive data. Just information that already exists inside the Canadian system — the kind of information that should, in any functioning democracy, be part of the public commons.
So I did what you’re supposed to do.
I emailed the Library of Parliament.
I asked my MP about it. The response was polite, but cautious. More or less: I know my limits. Let me know how it goes.
That told me something immediately.
There are lines here. Even elected officials are aware of them.
I filled out the web form. Sent another email. Then another. Reached out to the Speaker’s office.
Nothing meaningful came back.
So when I was in Ottawa last week, I decided to stop treating this like a digital process.
I went looking for a human being.
Spark Street
The actual Library of Parliament building is closed right now for renovations. But they have an office on Sparks Street.
So I went there. Walked up. Knocked on the door.
Which, apparently, is not something people do.
A security officer came over and asked what I was doing there. She was professional, composed, and — as it turned out — genuinely curious.
Fair question, honestly. I didn’t have an appointment. I didn’t have a badge. I had a laptop bag and a story.
I told her the truth. I run CanadaGPT. I’m trying to get access to data the Library manages. I’ve been through the forms and the emails. I just want to talk to someone.
To her enormous credit, she didn’t shut it down.
She listened. She asked questions. She wanted to help.
And then she did something that, in retrospect, was the most human moment of the whole trip: she walked me inside to ask a librarian on my behalf.
That part worked exactly how you’d hope it would.
A security officer — whose entire job is, in some sense, keeping people like me on the correct side of a door — decided that the correct thing to do was open it.
I want to hold onto that for a second.
Because it matters later.
The Librarian
I’m used to librarians being helpful.
They are, historically, one of the most radically generous professions we have. Their entire identity is built around helping people find information.
But this interaction felt different.
Less about the information.
More about me.
Who are you? Are you staff? Are you supposed to be here?
And underneath all of it, unspoken but unmistakable:
You’re not part of the system. And the system matters.
Eventually, the answer came. Only MPs and their staff have access to the Library.
That’s the rule.
What I got, in the end, was a piece of paper. An email address. A toll-free number.
The same channels I had already used.
That was the moment.
Not dramatic. Not confrontational.
But definitive.
That’s as far as you go.
What I Actually Wanted
I didn’t go there just for the data.
I went there to talk to a librarian.
Because librarians are the people who think about information as information. How it’s organized. How it’s catalogued. What gets indexed and what doesn’t. Which records sit next to which other records, and why. How people actually use a collection — what they come looking for, what they find instead, what they never find at all. The history of a dataset. Its silences. Its biases. Its quirks.
These are conversations I need to be having. These are conversations we need to be having.
Because right now, CanadaGPT has fifteen million nodes of indexed Canadian parliamentary and civic data. One question, one API, and you have access to the entire connected graph — Hansard, votes, bills, committees, lobbying, contributions, contracts, ATIP, federal organizations.
Imagine a country where you could ask: which companies lobbied on this bill, and how did MPs who took donations from them vote? — and actually get a real answer, sourced and checkable.
That’s the level this needs to operate at.
And I am not going to stop until I have all of it in there. Every public dataset. Every record of how this country governs itself. All of it, structured, connected, and available to anyone who wants to ask a question of it.
That is a significant thing to be doing. It deserves careful thought. It deserves the kind of conversation you can only have with someone who has spent their career thinking about how knowledge is shaped and shared.
And yet.
Everyone in this country right now is running around talking about how great AI is. Or how dangerous AI is. Both conversations are loud, and both are mostly surface.
What almost nobody is talking about is the actual substance underneath: how this is going to change the way we relate to our own information, our own institutions, our own record of ourselves.
That’s the conversation I wanted to have with a librarian.
Not can I have the data.
But what should someone like me know, before I do this at scale?
I walked into that office on Sparks Street hoping for a colleague. A thinking partner. Someone who would look at what I’m building and say here’s what you’re missing, here’s what to be careful about, here’s the thing people always underestimate about this collection.
Instead I got a phone number.
And I think that’s the loss, actually.
More than the access.
The loss is that the people best positioned to help us think about this are sitting behind a door, and the door is closed, and the conversation isn’t happening.
We are building a future and we’re not talking to each other about it.
The Five-Cent Voter
Over the past several months, a throughline has emerged across artificial intelligence that most policy conversations still have not caught up to.
Two Doors
Walking back down Sparks Street, I kept turning the afternoon over in my head.
Because the shape of it was strange.
A security officer — someone paid to be cautious — had been more willing to engage with me than a librarian, someone paid to share information.
That’s not a criticism of the librarian. She was following her rules, and her rules exist for reasons.
But the contrast stuck with me.
The officer treated me like a person who had shown up with a question.
The librarian treated me like a variable that didn’t match a schema.
And I don’t think either of them was being themselves, exactly. I think they were both operating inside a country that has forgotten how to be in a room together.
What Ottawa Feels Like Now
Here’s the part that’s harder to say, but I think it’s the real story.
We spent all of COVID locked away from each other.
And in Ottawa, that was doubled. First by the pandemic, and then by the convoy.
And all of that was the right call, at the time. We locked down to keep each other safe. That’s real.
But we haven’t come back.
We think we have, because the restrictions are gone and the masks are off and the cafés are full.
But psychologically we haven’t.
We’re still guarded. Still suspicious of proximity. Still surprised when a stranger knocks on a door.
The security officer who helped me — that was an echo of the Ottawa we used to have.
The librarian who couldn’t — that was the Ottawa we’re still stuck in.
We need to be talking to each other more right now.
Not less. Not through forms. Not through toll-free numbers.
We need the commons back.
The Commons
There’s a word we’ve quietly lost.
Commons.
It used to mean something very specific. Shared resources. Shared ownership. Shared responsibility.
Public knowledge was part of that.
Somewhere along the way, government started behaving like this information belongs to them.
It doesn’t.
It belongs to us.
And every form, every gatekept login, every “only MPs and their staff have access” is a small, quiet enclosure of something that was never supposed to be fenced off in the first place.
You can enclose a field. You can enclose a forest. You can also enclose a democracy, one access-controlled database at a time.
And you don’t notice it happening, because each individual fence looks reasonable.
Until one day you’re standing on Sparks Street with a piece of paper and a toll-free number, realizing that you’ve been perfectly, politely, professionally kept out of your own country’s record of itself.
Why I Started CanadaGPT
This is why I started CanadaGPT.
Not because I wanted to build a chatbot.
Because I wanted the commons back.
The data already exists. Hansard is public. Votes are public. Bills are public. Lobbying meetings are public. Political contributions are public. Federal contracts over ten thousand dollars are public.
It’s all technically there.
But “technically there” is not the same as available.
A PDF three clicks deep on a ministry website, behind a search box that doesn’t actually search, is not meaningfully public. It’s a performance of publicness.
Real access means structured, searchable, usable information.
It means the records talk to each other — that a vote connects to the bill it resolved, that the bill connects to the debates that shaped it, that the debates connect to the lobbying meetings that preceded them, that the meetings connect to the contributions that funded the campaigns of the people in the room.
It means an ordinary person can ask a question in plain language and get a real answer, sourced and checkable.
This is our data.
It belongs to the commons.
And we have the right to demand it.
All of us.
So Yes, In a Way
I got thrown out of the Library of Parliament.
Not physically.
Structurally.
And at the exact same time this is happening, the federal government is planning to reduce its workforce by thousands, including roles connected to information management.
We are entering the age of agentic AI — systems that can reason, retrieve, and synthesize information at scale — and the institutions that own the data are shrinking their capacity to manage and distribute it.
That’s not just inefficient.
It’s backwards.
But I don’t think the fix comes from inside those institutions anymore.
I think it comes from us.
From citizens deciding that the commons is worth rebuilding, and then building it.
I’m not dropping this. I’m not waiting for permission. I’m not accepting that this is just how it works.
If the data exists, it can be made accessible.
If it can be made accessible, it should be.
And if it should be, someone is going to do it.
We are that someone.
And we’re not waiting outside the door anymore.