The Crossing

EDO SEGAL: Here's how this works. For the next stretch I am barely here. You question each other directly — no moderator to soften it, no me to rescue anyone. Professor Zuboff, you've been waiting two centuries for this man to sit still long enough. You first.

ZUBOFF: Jeremy, I want to ask you the thing I've wanted to ask since I read your Panopticon letters as a young scholar. You designed a machine to make people good by making them feel watched. You knew — you said — that the watched would internalize the eye and govern themselves. How did you not see that a mechanism for installing an invisible watcher inside a person's head is the most dangerous gift you could ever hand to power? You trusted the watcher to be benevolent. Why? What in your study of actual power ever justified that trust?

BENTHAM: Because I did not trust the watcher, madam — I trusted the publicity. You have read my letters; read them again. I insisted, in the same breath as the tower, that the inspector be visible, the institution open, the accounts published, the whole apparatus subject to what I called the tribunal of public opinion. I never asked anyone to trust a benevolent watcher. I asked them to abolish the unwatched watcher by making him watchable. My error — and I see it now, in your century's light — was not that I trusted power. It was that I believed the publicity could be guaranteed by design, baked into the stone, when in fact the publicity is the first thing power strips out the moment it holds the tower. I built a machine whose safety depended on a feature — reciprocal visibility — that I treated as structural and that turned out to be merely optional, the easiest part to remove and the most profitable to remove. So I'll answer your question with the truth: I did see the danger. I named the safeguard. And I catastrophically underestimated how fast the safeguard would be discarded once the tower was built. Now let me ask you mine. You say abolish. I have listened all night and I still do not know what you mean concretely. When you say abolish the extraction — do you mean we may no longer build a model that predicts a heart attack from a heartbeat? Where, exactly, is your line between the prediction that saves the child and the prediction that farms her, given that you yourself admitted they run on one architecture? Draw it. Or admit it cannot be drawn, and that your "abolish" is as much a banner as my "reform."

ZUBOFF: Fair — you've earned a concrete answer and I'll give you one, and it's narrower than "abolish everything," because you've pushed me to sharpen it and that's what this hour is for. The line I draw is not at prediction. It's at the unilateral claiming of behavior as free raw material, and at the secondary market in predictions of persons. Here's the concrete version. A hospital may build a model on data its patients knowingly gave, governed by them, used only for their care, with the predictions never sold — that's the prediction that saves the child, and it survives, because the surplus was never claimed unilaterally and the future was never put up for sale. What dies is this: the claiming of behavioral surplus as a byproduct of a service, without genuine consent, and the sale of prediction products about people in markets they're not party to. The heartbeat model in the hospital lives. The heartbeat sold to an insurer to price her out dies. Same data, different regime of ownership and sale. So my line isn't between predictions — it's between governed prediction in the service of the predicted and extracted prediction sold against them. You asked where the line is. It's at consent and the prohibition of the futures market. And Jeremy — I'll grant you something in return. That line, drawn that way, looks more like your "reform" than my rhetoric of "abolition" usually admits. I abolish the extractive mechanism. I keep the governed instrument. Maybe we've been closer than two hours of fighting suggested.

BENTHAM: ... Then I must ask the question that closeness forces, and ask it without flinching, because if we are closer than we seemed, the remaining gap is the one that matters most. If you keep the governed instrument and I bind the extractive one — if your hospital model lives and my forbidden-sale clause stands — then we have, in substance, the same architecture: prediction permitted under reciprocal, consented, non-commercial governance; prediction forbidden where it is extracted and sold. You call the arrival "abolition of surveillance capitalism." I call it "the Panopticon with its windows restored." Here is my real question, the one under all the others: is our remaining disagreement substantive, or is it only temperamental — you believing the windows can never be made to stay, and I believing they can? Because if it is only that, then the whole fight between reform and abolition reduces to a single empirical wager about human institutions — whether publicity, once restored, can be kept. And that is a wager, not a principle. Tell me honestly: is anything left between us but the question of whether the windows hold?

ZUBOFF: ... That's the most honest question anyone has asked me about my life's work, and I'm going to answer it honestly even though it costs me. Yes. I think — after tonight — that what's left between us is largely the wager on whether the windows hold. You believe accountability, once legislated, can be sustained against the gravity of profit. I believe the gravity of profit has, every single time in my forty years of evidence, torn the windows back out faster than law could install them — and that at some point a remedy that has failed every time it's been tried stops being reform and becomes an alibi. But I won't pretend that's a difference of principle. It's a difference of historical faith. You have two hundred years of conviction that institutions can be made good. I have forty years of data that this particular institution metabolizes every reform and accelerates. Neither of us can prove the future. So here's my concession, and it's real: if you could guarantee the windows hold — genuinely guarantee reciprocity, consent, the forbidden futures market, sustained against profit — then I would keep your instrument and call off the abolition, and we'd have no quarrel left. My abolitionism isn't a love of destruction. It's a judgment that the guarantee is impossible. And I notice that's an empirical judgment, not a moral absolute, which means — God help me, Jeremy — you've moved me about a degree, and a degree is a lot, across two centuries.

EDO SEGAL: I'm coming back into the room now, because the reader can't see your faces and I have to mark what just happened: that was the first exchange of the night where neither of you was trying to win. You found the floor of your disagreement, and it's smaller and harder than either of you walked in believing — not a clash of principles but a wager about whether human institutions can hold a window open against the pull of profit. That's the gift of three hours. Hold it. One round left before the close. The deepest one.