I don’t know how much of it is private. I don’t know what he opens, what he reads, what he keeps. I don’t know who else has that address filed next to my name, or in how many systems, or for how long. What I know is that there is a version of me living inside other people’s records: forwarding addresses, account notes, old contact fields that were never updated and I have no way to see her, let alone correct her. She is out there making an impression on my behalf. She arrives before I do.

I used to think the thing I was afraid of was being exposed. The recurring nightmare a lot of people have, the one where you’re on stage in your underwear. I had a version of that. But the longer I spend building SOLO, the more I think I had the fear backwards. It isn’t the fear of being seen. It’s the fear of being seen wrong, by someone with the power to make a decision that could impact your life, in a file you’ll never get to read. We built a $50 billion-a-year data industry to decide what people can access and never gave them a mirror.

A year ago, I was on a different kind of stage, at a town hall, speaking with , who was way ahead of his time in understanding how AI would reshape ordinary economic life.

The room was full of bankers, fintech leaders, policymakers, and consumer advocates. The conversation turned, as it always does, to consumer data and privacy. And I said very casually the thing my team had spent months hoping I would never say in public:

“I’m not sure I care that much about privacy. I kind of like that Instagram knows my style. It means I can just buy the dress.”

Not my finest moment, as the founder of a consumer reporting agency. But I meant it, and I still mean it, so let me say it again, more carefully.

I don’t mind that a restaurant knows I’m gluten-free. I want them to know. I want them to tell the next restaurant, honestly, because then I get shown things I can actually eat, and because if I order something dangerous I’d like someone in the room to know it could hurt me. What I want is to know that they know. I want to be able to change it when it changes, when I decide I’m done being gluten-free and resent the whole category. And I want it used for the reason I handed it over, not silently repurposed into something that gets used against me when I’m not looking.

That last part is the whole point. That’s the line nobody draws.

The invasion isn’t that my information exists. The invasion is secondary use: data I gave for one purpose, quietly turned into another. You log into your bank account once to make a payment because its the only method available at checkout. Later, the income data from that same account becomes the reason a landlord denies your application, and no one asked you, and you never find out which fact did it. You logged in to pay. You didn’t even know you were sharing your income. Now its an input to a decision being made about your life. You cannot supplement it. You cannot even see it.

We keep calling that a privacy problem. It isn’t. Privacy is about whether you are seen. This is about whether you have any say in how you’re used. That is agency, and agency is the graver thing to lose. You can survive being seen. It is much harder to survive being decided.

I learned the difference in a field where the stakes were never abstract.

I started my career as a data scientist in cyber threat intelligence, in the parts of the internet where harm organizes itself: the deep and dark web. I worked cases involving children recruited into terrorist networks through gaming apps. I traced money laundering through financial systems. It was work where being wrong about a person had consequences, so the field had a discipline the rest of the economy has simply chosen not to have.

The discipline was this: a model can tell you where to look. It cannot tell you what’s true.

In that world, prediction had a real but limited job. It surfaced patterns, filtered noise, pointed your attention. But you could not prosecute someone because a system found them suspicious. A signal was a place to start, not a verdict to act on. Once you had a suspect, the work shifted, toward evidence you could corroborate, toward records someone could confront and contest, toward a case that would survive scrutiny. The standard was never “the model thinks so.” The standard was whether it would hold up in a fair process.

Prediction is not proof. Risk is not guilt. Correlation is not evidence. A score is not a verdict.

And then I looked at consumer life and watched every one of those lines get erased. Not because anyone decided they didn’t matter, but because the stakes got reframed as operational instead of punitive. Nobody is marched into court when an apartment is denied, or a credit limit quietly cut, or a job application killed by a screening filter before a human ever reads the name. So the governance got looser, the scrutiny lighter, the standard weaker. We tolerate things inside private systems we would never tolerate in a courtroom. But for the person being flattened into a risk profile, the courtroom distinction means nothing. The apartment is still gone.

The thing people picture when they hear this is Nosedive, the Black Mirror episode where one visible score governs your whole life. I think the real version is stranger and harder to fight, because there is no single number to point at. Instead there are a thousand of them, each sitting in its own system: a fraud flag here, a tenant screen there, an identity check, an internal risk note, an AI summary that knows your patterns but not your context. None of them announces itself. They show up as outcomes. A delay. A denial. A request to prove the same thing again. A silence where an approval should have been.

This is what ties together the problems we insist on discussing separately. Credit bureaus decide your housing. The internet decides your reputation. AI will increasingly decide your work. We treat those as three different topics because they live in three different industries. They are the same topic. All three run on information systems making decisions about people, with almost no accountability and almost no way to correct the inputs. We regulate the bad outcomes. We have never regulated the architecture that produces them.

When I first started this company, I had a line I used in every room. A credit report is like being sued without being served. I meant it as an accusation. The bureaus hold a file on you, a file that decides whether you get the apartment, the loan, the rate, the second look, and then they charge you something like twenty four dollars a month for the privilege of reading it. I thought that was the most insulting thing I could imagine doing to a person: making them pay to see the case being built against them, in their absence, by institutions claiming to value their relationship.

I want to tell you how naive that version of me was.

Because today, twenty four dollars to see what is held on you sounds like a bargain. It sounds like a right. We have moved into an era where, for most of the decisions that actually shape your life, you cannot even pay to see the underlying records. You do not get the bill and the file. You just get the outcome. The data goes out, gets recombined, gets sold, gets fed into a model, and the version of you it produces goes to work in rooms you will never enter, and there is no twenty-four-dollar door, no monthly subscription, no customer-service line. The thing I once called extortion turns out to have been the last functioning correction mechanism we built. At least it let you in.

So here is what I’ve come to believe the fix is not. It’s not a “fair score.” It is a record you can argue with.

A score asks to be believed. A record can be inspected. A score compresses context until there is nothing left to dispute. A record can preserve it: who made the claim, what evidence they had, what the other side said, what was verified, what is still contested, what got corrected, and what the whole thing is even allowed to be used for. Truth, in the only version I find livable, looks less like a verdict handed down from behind a curtain and more like a governed argument. It is the thing that survives a fair process. It is adversarial, and that is not a flaw. That is what makes it real.

We live with the consequences of not having a record every day, informally and exhaustingly. We Google before dates. We ask for screenshots. We read the LinkedIn, inspect the Instagram, save the receipts, search the name before we send the money. We have all become private investigators because the systems meant to protect us move too slowly and refuse to show their work.

The nightmare was never that some AI would issue one dramatic ruling over your life. It’s quieter than that. It’s that judgment becomes ambient, that thousands of small, invisible decisions get made about you, all the time, everywhere, by systems you were never meant to understand and never given a way to answer.

The right that matters in this era is not the right to be seen. We’re all already seen, constantly, in files we’ll never open.

It’s the right to see how you’re seen.

Right now, we don’t have a mirror. My ex still gets my mail, and somewhere a version of me is making my case in a language I never agreed to, before people I will never meet (and losing).

I would simply like to be in the room when decisions about me are being made. I think you would too.