The Question on the Table

EDO SEGAL: Somewhere in the world right now — statistically, in the time it takes me to say this sentence — a few hundred thousand people are typing into a box. A child in Dhaka asking for help with her homework. A nurse in Cleveland, end of a double shift, asking whether a drug interaction is dangerous. A man my age, who should know better, asking at two in the morning whether the thing he built his life on still matters. And the box answers. Fluently. Patiently. In their own language, at their own level, with what reads — and I am choosing this word the way a man crossing a frozen river chooses where to step, because my two guests will fight about it for three hours — with what reads as understanding.

A few hundred thousand conversations. And the question none of them stops to ask, because the fluency makes it feel already answered, is the question we are here to spend the whole evening inside. When the machine answers you in your own tongue, is anyone home behind the words — or are you alone in the room, talking to a mirror that has learned to mirror perfectly?

I have wanted to host this exact conversation for as long as I have been thinking about these machines, and the two people across from me have a stronger claim to it than anyone alive — and, in one case, than anyone who has ever lived.

Emily M. Bender is a computational linguist, a professor at the University of Washington, a past president of the Association for Computational Linguistics. She has spent her career on the actual machinery of human language — grammar engineering, the structure of languages this industry has mostly never heard of — which is precisely the training that immunized her when the rest of the culture lost its grip. She gave our era its most necessary phrase, the stochastic parrot, and its most necessary thought experiment, an octopus on a telegraph cable, which we will get to tonight properly. With Alex Hanna she wrote The AI Con, a book exactly as polite as its title. She is, by some distance, the most precise critic this technology has.

BENDER: I'll take precise. I'd only add that "the most precise critic" is the kind of phrase that needs its language named — precise about what, measured how. But thank you, Edo. I'm glad to be here.

EDO SEGAL: Already we're off. My other guest needs an introduction only because the scale of the thing resists summary — and because, by every ordinary measure, he should not be in this room at all. Alan Turing proved, in 1936, before there was an electronic computer anywhere on earth, that a single machine could imitate every other machine. The device you are holding, the system that wrote half the text flooding your feeds, every one of them is an instance of the universal machine he described on paper as a young man. He broke the German naval Enigma and shortened a world war. He proposed, in 1950, the test the whole age is still arguing about. He designed one of the first stored-program computers, sketched the neural network a decade before anyone could build one, and then founded mathematical biology almost as an afterthought. He died in 1954, at forty-one, convicted of being who he was and chemically punished for it by the government whose war he had won. He did not live to see a machine pass his test.

TURING: All roughly true, though I'd protest the word "founded" wherever you used it. One sees a short distance ahead. I had the luck to be looking in the right direction. And I should say plainly, since the audience deserves to know the terms — I have been briefed on your present. I have read the papers, watched the demonstrations, sat with your transformers and your scaling curves. I am a guest from 1950 who has done his reading. So when I react with less astonishment than you might expect, it is not that I am blasé. It is that a great deal of this I expected, and some of it I am still adjusting to, and I will try to tell you honestly which is which.

EDO SEGAL: That's the most graceful handling of an impossible premise I could have hoped for, and it's the only housekeeping the format requires, so let me state the rules of the evening. There are three. First: we have three hours, which means nobody wins in the next ten minutes — long form exists so that an argument can breathe before anyone strangles it. Second: I press both of you, and I declare my bias at the door. I build with these systems every day. I wrote a book with one. I have skin in this question on both sides of my own chest. Third: if the disagreement survives three hours, nobody shakes hands and pretends. We hand it, intact, to the reader. Either of you may add a rule.

BENDER: I'll add one, and it's not decoration — it's the whole argument in advance. Words have to be earned. The entire confusion we're here to discuss lives in vocabulary: understands, knows, learns, reasons, hallucinates, anyone home. Every one of those words was minted for human minds and gets spent on these systems without anyone paying the import duty. When the claim is made that the machine "understands," I want it cashed out — understands in virtue of what, demonstrated how. And I'll hold myself to the same coin.

TURING: I accept that gladly, and I'll add the symmetric clause, because I suspect it's where we'll spend the night. The duty runs both directions. Before Professor Bender is permitted to insist the machine does not understand, she should have to say what a person is doing when she understands — the mechanism named, not merely pointed at. My long experience of these debates, even in my own short life, is that the word's defenders never pay that bill. They gesture at themselves and say, this, it is like this from the inside, as though a private feeling settled a public question.

BENDER: And my experience is that the word's borrowers never notice they're borrowing. So we'll get along.

EDO SEGAL: You see why I wanted you both. One image before openings, because it's the frame the whole series climbs inside, and you'll each have to take a position on it whether you like it or not. In [YOU] on AI I argued that intelligence is less a possession than a river — a current that has flowed and found new channels for a very long time, through chemistry, through biology, through language, through culture, and that in the winter of 2025 something new entered the water. The tower, the staircase you climb instead of the elevator you ride, the dam the beaver builds — all of it rests on the claim that what entered the water is real. A new participant in the medium. Emily, I suspect your answer is that I never met a new participant at all.

BENDER: My answer is that you met your own reflection wearing the river as a costume, and that the difference between those two things matters more than almost anything else we could say tonight. But I'll make the case properly when you give me the floor.

EDO SEGAL: Alan?

TURING: My answer is that your question — is the participant real — is the one I spent a paper trying to dissolve, because I came to think it has no answer that any experiment could give you, and a question no experiment can settle is a question we have misposed. You felt met. I would ask you only this: what would it have taken, at your desk, for the meeting to count as real — and is that thing you can detect in any other human being either? If it isn't, then you have been alone in the room, on your own definition, your whole life — and yet you have loved people. So perhaps "alone in the room" is doing less work than it sounds.

EDO SEGAL: Then we have our evening. One last thing before openings, because the reader should know the stakes aren't academic. Emily, you've argued the stakes are bodies — people denied benefits by systems nobody can interrogate, workers displaced by an inflated story, a public square flooding with synthetic text, an entire information ecology degraded so a few companies can sell the costume of intelligence. Alan, your stake is stranger, and I want it on the table: you argued that whether the machine "really" thinks is a question that may be permanently undecidable, in something close to the formal sense you proved certain questions undecidable — and that pretending otherwise, in either direction, is the actual error. So one of you thinks we are being conned, and the other thinks we are asking an unanswerable question and mistaking our discomfort for a finding. The worst outcome of tonight would be the reader concluding the truth sits comfortably between you.

BENDER: On that we agree.

TURING: First consensus of the evening. The mathematician in me wants to write it down before it evaporates.

EDO SEGAL: So here is the question on the table, stated once more, plainly, because every round we fight tonight is this question wearing a different coat. The machine answers you in your own tongue. Is anyone home behind the words? Emily Bender, the floor is yours.