Is AI the Real Thief? Signatures, Bullets, and the Trust Nobody Reported Stolen
Introduction When a thief uses a crowbar to break into a house, do we arrest the crowbar? It sounds like a silly question. Of course we don't. We arrest the thief. The crowbar goes into an evidence bag and, presumably, into a very dull retirement. And yet, when it comes to artificial intelligence, a great many of us are standing in the street shouting at the crowbar. "AI is stealing our jobs!" "AI is stealing our art!" "AI is stealing our words!" And some of that is true, or partly true, and we'll get to it. But I've been noticing something. When an illustrator who has spent fifteen years mastering her craft uses AI to speed up her work, nobody accuses her of theft. She's still the illustrator. She just has a faster brush. The same goes for the composer, the programmer, the teacher, the translator who uses these tools to do more of what they already know how to do. So why do some uses feel like robbery and others don't? And if there really is a thief in this story, who is it? The machine? The companies that built it? The man who has never written a paragraph in his life and published sixty books last month on psychology, medieval cheese, and the private lives of the Victorians, none of which he has read? Or, and this is the one nobody likes, is it us? That's what this episode is about. We'll look at all the suspects, one by one, like a detective in an old novel gathering everyone in the library. We'll talk about bullets and bridges and used cars, and an old Latin word that I think holds the key to the whole case. Here's how we'll do it. First, a story: a composer, a blackout, six flights of stairs, and a room full of glowing phones. Then you and I will talk, laughing where we can, until we reach the part where laughing doesn't fit anymore. And at the end, I'll leave you with some questions. So let's go to Beirut. The generator is humming, and someone's phones are listening to music nobody wrote. The Story Nour Khalil found out where her royalties were going on a Sunday, over stuffed grape leaves, from her cousin's husband, who didn't know he was confessing. "Music," Rabih was saying, waving a fork. "Music is the future, Nour. You of all people should know that. I've got a label now." "A label," Nour said. "Forty-two artists." He grinned. "Well. Forty-two names. Lo-fi, ambient, sleep sounds, rain on a tin roof, that kind of thing. People love rain on a tin roof. Very relaxing. Streams all night." Nour was thirty-three. She had been composing since she was nine, when her grandmother put her in front of an old upright piano and she refused to get up for dinner. She had a degree from the conservatory and a master's from a university in Lyon that she was still, technically, paying for. She had released four albums of piano and string music, mostly on her own. Her most-streamed track had eighty-one thousand plays. Last month, from every streaming platform on earth combined, she had earned thirty-eight dollars and forty-one cents. "Who composes it?" she asked. Rabih laughed and tapped his temple. "The machine. I just tell it what I want. 'Sad piano, rain, a bit of cello.' Two minutes, done. Then another one. And another one." He leaned over. "You should see my streams, Nour. Honestly. You should come see the setup." His wife, Nour's cousin Lina, was not laughing. She was looking at her plate. That night, at 11:20, Lina called. "Can you come? Please. He's in the flat in Dekwaneh. He's been there since dinner. He's got the generator bill here on the table, nine hundred dollars, Nour. Nine hundred. For one month. He says it's for his servers." Her voice cracked. "I don't even know what he's doing. You understand music. Can you go talk to him?" Nour didn't really want to. But she knew Lina had been carrying the rent on her teacher's salary since Rabih lost his job at the bank in 2020, when the whole banking system froze and people's savings vanished behind glass doors that no longer opened. She knew Rabih had been one of the IT men who kept those very systems running. She knew he'd come home one day with a cardboard box and not spoken for a week. She took a taxi to Dekwaneh. The building was one of those half-finished concrete towers that had been waiting for its final coat of paint since before she was born. The power was out on the whole street. The only light came from the generator shops, their cables looping across the road like black vines, and from one window on the sixth floor, which was glowing a strange, flickering blue. The elevator was dead, of course. She climbed six flights of stairs by the light of her phone, her breath loud in the stairwell, a cat watching her from the fourth-floor landing with profound disapproval. The door to the flat was open a crack. She pushed it. For a moment, she simply stood there. The living room had been emptied of furniture. In its place, along every wall, stood metal shelving, the kind you'd find in a warehouse, and on every shelf, in neat rows, sat phones. Hundreds of them. Old phones, cracked phones, cheap phones, each one plugged into a thick tangle of chargers, each one glowing with a screen showing a music app, each one playing, silently, through no speaker at all, a song. The room hummed with fans. It was warm as an oven. The screens flickered in waves, blue and white, like a school of fish turning in dark water. In the middle of it all, at a folding table with two laptops and a half-eaten man'oushe, sat Rabih. "Nour!" He seemed genuinely delighted. "You came! Welcome to Callous Records." He spread his arms. "Four hundred and sixty phones. Each one's a listener. Each one listens to my artists all night. The platform pays per stream. Nobody's hurt. Beautiful, no?" "Nobody's hurt," she repeated. "Who's hurt? I'm not stealing from you. I'm stealing from the big platform." He shrugged. "They're a billion-dollar company. Believe me, they won't miss it." She walked slowly along the shelves. One phone was playing Rain on Tin Roof, Vol. 17. The next was playing Ambient Study Focus Deep Calm. The artist names scrolled past: Velvet Hours, Lunar Tide, The Quiet Room. "Rabih," she said. "Do you know how streaming money works?" "Sure. Streams equal money." "No. It's a pool." She turned around. "The platform takes all the money from all the subscribers in a country, puts it in one big pot, and divides it by the total number of streams. Every stream gets a slice. So when your four hundred and sixty phones play rain on a tin roof all night, the pot doesn't get any bigger. The slices just get smaller. For everyone. Your money isn't coming from the platform. It's coming out of the same pot as mine." She pulled out her phone and showed him the statement. "Thirty-eight dollars. Some of that should have been forty." Rabih looked at the number for a long moment. Something moved across his face. Then he laughed, not unkindly. "Two dollars, Nour. You came all the way up here in the dark for two dollars?" "It's not two dollars. It's two dollars from me, and two from the next person, and the next, and the next. Everybody who ever sat at a piano. And it's not even the money. It's..." She searched for it. "People search for piano music, and they get this. They get Velvet Hours. And they think that's what piano music is." "And what did the banks take from me?" His voice changed. It went hard and quiet. "My savings. My father's savings. Everything. They didn't ask. They just closed the doors. Nobody went to prison, Nour. Nobody. Twenty years I kept their systems running. So don't come here and talk to me about who is a thief." The fans hummed. On the shelf behind him, a phone flickered and restarted itself. And then his laptop pinged. He glanced at it, and all the color went out of his face. "No, no, no." He was typing fast. "They've flagged my distributor account. They're doing a purge. Tonight. At midnight they take everything down." He looked at the clock in the corner of the screen. 11:51. "Four years of catalog, Nour. Everything." He turned to her. And she saw an idea arrive in his eyes like a car's headlights sweeping across a wall. "You have a verified account. A real artist account. Four albums. Clean history." He was on his feet now. "Let me upload the best tracks under your name. Just the top five hundred. As collaborations, remixes, whatever. They'd never flag you. I'll give you thirty percent. No, forty. Forty percent, Nour. You'd make more in one month than in your whole career." "You want to put my name on music nobody wrote." "I want to put your name on music people actually listen to." He said it gently, almost kindly, and that was the worst part. "Thirty-eight dollars, Nour. How long are you going to keep playing for thirty-eight dollars?" 11:54. She stood in the blue light, surrounded by four hundred and sixty phones listening to nothing, and thought about her grandmother's piano . She thought about the master's degree she was still paying for. She thought about her rent, and the dentist she'd been putting off, and the eighty-one thousand people who had listened to her best track, some of whom had written to her to say it got them through a hospital stay, a divorce, a first winter in a strange country. She thought about her name. N-O-U-R. Light. 11:57. "No," she said. He looked at her as if she'd said something in a language he didn't speak. "The offer stands until morning," he said quietly. "In case you change your mind. I'll send you the details." At 12:00, the laptop chimed. Rabih didn't look at it. The phones on the shelves kept playing. Some of them, she noticed, were already showing error messages. Others weren't. Not yet. She went back down the six flights in the dark. The cat on the fourth floor had gone. At the bottom, in the street, the generator shops were still roaring, and the cables still looped overhead, carrying electricity bought by the hour to people who could afford it. She stood on the curb waiting for a taxi and looked at her phone. A new message from Rabih. The details, as promised. Account names. Percentages. A link. Under it, the notification from her own streaming dashboard, the one she checked every night before sleep, like a prayer: Your monthly earnings: $38.41. She didn't answer the message. She didn't delete it, either. Reflection and Discussion Alright! Let’s reflect on the story and dive deeper in our topic for today. Let's begin with a working day in the life of a modern author. Not the old kind, the kind who sits by a window agonizing over a comma. The new kind. He wakes up at nine. He opens one AI tool and types: "Give me twenty profitable nonfiction book ideas." It obliges. Among them: The Psychology of Everyday Decisions, A History of Cheese-Making in Medieval Europe, and Scandal and Desire in Victorian England. He pastes those into a second AI tool, which writes the books. A third tool designs the covers, each featuring a slightly haunted-looking stock photograph and a font that suggests authority. By lunchtime, he has published twenty books. By dinner, another twenty. He has not read a single one. He could not tell you what's in chapter three of the cheese book if you held a wheel of Gruyère over his head. And his social media bio now says: "Author of 400+ books. Thought leader. Storyteller." Tolstoy took about six years to write War and Peace. This man would consider that a disappointing morning. And here's the detail I love most. In September 2023, Amazon, which runs the largest self-publishing platform in the world, quietly announced a new limit. Self-published authors could now publish no more than three new books per day. Three. Per day. The fact that the company felt the need to set this limit tells you everything about what was happening. Somebody, somewhere, had been publishing enough books every day to make a trillion-dollar company nervous. Three books a day is still, by any historical standard, a staggering pace. Agatha Christie wrote sixty-six detective novels in her whole life, and she was considered astonishingly prolific. Under Amazon's generous limit, you could overtake her in twenty-two days, with weekends off. Then there's the problem of whose name goes on the books. In the summer of 2023, the American writer and publishing expert Jane Friedman discovered a handful of books for sale under her name that she had never written. They were bland, generic guides to writing and publishing, exactly the sort of thing she might plausibly have written, and they appeared to have been generated by AI and attached to her name to trade on her reputation. She complained, and at first, she said, it wasn't easy to get them removed, because she couldn't prove the name was a registered trademark. Her own name. Think about that. A writer had to argue with a platform about whether she owned the name she'd spent her career building. Eventually, after a public outcry, the books came down. But how many writers have that kind of public voice? And it isn't just books. In 2024, an American named Michael Smith was charged with using AI to generate hundreds of thousands of songs and then using bots to stream them billions of times, collecting more than ten million dollars in royalties between 2017 and 2024. In March of this year, he pleaded guilty to conspiracy to commit wire fraud and agreed to forfeit more than eight million dollars. And I have to share the names of some of his artists, according to the prosecutors, because no novelist could invent them. There was a band called Calorie Screams. There was one called Calvinistic Dust. There was Callous Post. One of the songs was called "Zygotic Washstands." I've read that list several times now, and I still can't decide whether it's the saddest or the funniest thing I've ever seen. Somewhere, a machine was asked to invent a band name, and it said, with total confidence, Calvinistic Dust. And then there are the instant programmers. In early 2025, the AI researcher Andrej Karpathy coined a phrase that spread across the internet in about a week: "vibe coding." He meant, half jokingly, a way of building software where you describe what you want to an AI, accept whatever code it gives you, and don't really look at it too closely, and he was talking about quick weekend projects. The phrase was charming. What followed was less so. Suddenly, people who had spent their entire lives avoiding computers, people who still print their emails, were announcing themselves as "founders" of apps. Apps that store your password in a place roughly as secure as a sticky note on a café window. Apps that work beautifully until the second user signs up. I have nothing against a weekend project; I've built a few terrible ones myself. But there's a difference between tinkering in your garage and inviting the public to drive your homemade car on the highway, with their credit card details in the glove compartment. The prosecutor put it well. The songs were fake, and the listeners were fake, but the money was real, and it came out of the pockets of real musicians. That's exactly what Nour was trying to tell her cousin's husband in the story. Streaming royalties, on the major platforms, come out of a shared pool. Every fake stream doesn't make the pot bigger. It makes everyone else's slice smaller. And of course, in every gold rush, the people who get rich most reliably are the ones selling shovels. So alongside the instant authors, there's now a thriving second industry: people selling courses on how to become an instant author. "Make ten thousand dollars a month with AI books. No writing required!" The videos usually feature a young man in a rented sports car, standing in front of a rented villa, explaining that he quit his job to publish books he hasn't read, and that for just four hundred and ninety-seven dollars, you can learn his secret. His secret, as far as I can tell, is selling the course. It's a beautiful little ecosystem: machines writing books nobody reads, promoted by courses taught by people who've never written books, bought by people hoping to never write books either. At no point in the entire chain does anyone have to read anything. It's the first literary movement in history built entirely on the avoidance of literature. Now, compare all of that to the illustrator I mentioned at the beginning. She's spent fifteen years learning anatomy, color, composition, the way light falls on a cheekbone at five in the afternoon. Now she uses AI to generate rough backgrounds, or to test ten color palettes before breakfast, and then she paints over it, corrects it, throws most of it away, and signs only what she's satisfied with. Why doesn't anyone feel robbed by her? I think it's because every image that leaves her studio has passed through her eyes and her judgment. The machine made her faster. It didn't make her absent. She's still the one standing behind the work. Hold on to that idea. It turns out to be the key to the whole case. So, clearly, something is being stolen. Now let's try to figure out who the thief is. Let's gather the suspects in the library. You know the scene. It's raining outside. The fire is crackling. The detective, in a slightly too-tight waistcoat, paces in front of the assembled guests. There's the machine, sitting very still in the corner, humming faintly, looking suspiciously innocent. There's the company, in an expensive suit, checking its phone and asking whether its lawyers can join by video. There's the instant author, holding a stack of four hundred books, none of which he has opened. And there, at the back, trying very hard to look like part of the wallpaper, is a fourth figure, wearing headphones and holding a ninety-nine-cent e-book. We'll get to that one. The detective always saves the most awkward suspect for last. The first suspect is the obvious one: the AI itself. The machine that writes the books, composes the rain on the tin roof, and names the band Calvinistic Dust. The case for the defense is the one you'll hear most often, and it's a good one. A tool is a tool. The same AI that lets a fraudster publish four hundred books lets a dyslexic student finally write the essay that's been trapped in her head. The same software that composes fake ambient music helps a professional composer sketch orchestral arrangements in an afternoon instead of a week. It's the old argument about weapons: a bullet can be fired by a soldier defending a village or by a murderer in an alley. The bullet doesn't choose. Blame the hand, not the bullet. I find that argument mostly persuasive. But a critical thinker should push on it, because it's not the whole story. In 1980, the political theorist Langdon Winner wrote a famous essay called "Do Artifacts Have Politics?" His argument was that tools and technologies are never completely neutral. The way they're designed makes some things easy and other things hard, and that shapes what people do with them. His most famous example was the parkways of Long Island, near New York. He argued that the planner Robert Moses deliberately built the bridges over these roads unusually low, so that public buses couldn't pass under them, which in practice kept poorer people, many of them Black and Puerto Rican, who relied on buses, away from the beaches. The bridges themselves, Winner suggested, had politics built into their concrete. It's a brilliant example. It's in countless university courses. And here's the twist, which I love because it's so on theme: it's contested. In 1999, a sociologist named Bernward Joerges published a critique arguing that the evidence for Moses's intent was thin, that commercial vehicles, including buses, were banned from those parkways anyway, and that there were other routes to the beaches. So the most famous parable about how tools carry intentions may itself carry more intention than evidence. But here's what survives the correction, and it matters for our case. Whether or not Robert Moses meant it, design choices do make some behaviors cheaper and others more expensive. And AI has made one thing almost free: volume. Before these tools, if you wanted to publish four hundred terrible books, you'd have had to pay four hundred terrible writers. The tool didn't create the desire to flood the market. But it removed the price tag. And when you remove the price of something, you get a lot more of it. That's not a moral argument. It's just economics. So the bullet may not be guilty. But a factory that makes bullets free, fast, and unlimited is not quite the same thing as a single bullet. Which brings us to the second suspect: the companies that build these tools. Here, the questions are serious, and I want to be fair, which means naming facts that some people will find uncomfortable in different directions. Many AI companies trained their systems on enormous amounts of text, images, and music created by human beings, often without asking and without paying. Dozens of lawsuits are working their way through the courts over exactly this. The first major settlement came in one of them: a group of authors sued the AI company Anthropic, which makes the Claude models. In a ruling last summer, a federal judge found that using books to train AI could count as "fair use" under American copyright law, which was a significant win for the company. But the same judge ruled that downloading millions of books from pirate websites to build a library was not protected. The company settled for one and a half billion dollars, roughly three thousand dollars per book, and in July of this year, a judge approved that settlement, which the authors' lawyers called the largest known copyright recovery in history. Notice how complicated that is. The court said that learning from books wasn't theft. It said that taking them from pirates was. Some people will read that and say the companies are thieves who got caught. Others will say the system worked: the courts drew a line, and the company paid for crossing it. Many other cases, against many other companies, are still undecided, and courts in different countries may draw very different lines. A critical thinker doesn't have to settle all of that tonight. But it's clear that the companies aren't innocent bystanders either. They make choices about what data to take, what products to build, and how to market them. And when a company advertises that you can "write a book in a day," it isn't selling a faster brush to illustrators. It's selling a shortcut to people who never wanted to paint. And we shouldn't forget the platforms, the marketplaces and streaming services where all this flood ends up, because they're a sort of accomplice by business model. A bookstore that earns a commission on every sale doesn't lose much sleep over whether the sale was a lemon. When Amazon introduced its three-books-a-day limit, it also began asking self-publishers to tell the company whether their content was AI-generated. Notice the direction of that information, though. The publisher tells the platform. As far as I could find, the platform didn't show that information to the readers browsing the shelves. So the one party that most needs to know, the person about to spend money and hours on the book, is the one party kept in the dark. It's a bit like a restaurant asking its chef whether the fish is fresh, writing down the answer, and then putting the notebook in a drawer. Now, the third suspect, the one this whole episode began with. The instant author. The man with four hundred books he hasn't read. The instant composer with forty-two fake artists. The person who has never shown the slightest interest in writing, music, or programming, and who suddenly, the moment the tools appeared, decided to become all three. I'll admit, when I first thought about this episode, this was my prime suspect. The person who never cared about the craft, never struggled with it, never spent a single night trying to fix a paragraph, and now floods the shelves with things he's never read, and calls himself a writer. But then I read the details of the Michael Smith case again, and something stopped me. Michael Smith was a musician. He'd been uploading his own music to the platforms before the scheme began, according to the indictment, and he found that his own catalog wasn't big enough to earn much. So he turned to AI and bots. He was not someone who had never shown any interest in music. He was, in a sense, one of the people this episode is supposed to be defending. That's an important complication, and I think it forces us to refine the argument. The line isn't really between people who have a lifelong passion and people who don't. Plenty of passionate people behave badly, and plenty of newcomers behave well. The line is somewhere else. And there's another complication. Long before AI, human beings were already doing this. In the late 2000s, companies called content farms paid writers a few dollars an article to churn out enormous numbers of low-quality web pages designed to catch search traffic. How to tie a tie. How to boil an egg. How to cure a hangover, in eleven slightly different versions. One of the biggest of these companies was producing thousands of articles a day. It became such a problem that in 2011 Google changed its search algorithm specifically to push them down. So the flood isn't new. The thief was already in the house. AI just handed him a much bigger bucket. So let me try to state it more precisely. The thief isn't the person who comes late to a craft. It isn't the person who uses AI. It's the person who puts his name on work he hasn't read, hasn't checked, hasn't cared about, and offers it to people who will trust it. It's the person who extracts value from a shared space while contributing nothing to it and deceiving the people who use it. And I want to be careful here, because I think my original instinct had a little gatekeeping hidden in it, and I should own that. It's tempting to say, "You never showed any interest in writing, so you have no right to write now." But everybody starts somewhere. Laura Ingalls Wilder published her first Little House book in her sixties. Plenty of people discover late in life that they have something to say. There are people with disabilities who can finally express themselves with these tools. There are people writing in a second or third language, many of whom listen to this show, for whom AI is the ramp that gets them to the door. If a retired engineer in Tripoli decides at sixty-five to write a book about his father's village, uses AI to help him shape it, reads every word, argues with every sentence, and puts his name on it because it's genuinely his, I'm not going to call him a thief. I'm going to buy his book. The difference between him and the man with four hundred books isn't talent, or history, or passion. It's whether there's a human being standing behind the work, having read it, cared about it, and meant it. So why does it matter so much? Why not just let the four hundred books sit there, unread, harming nobody? Here's where an economist can help us. In 1970, George Akerlof published a paper with a wonderfully unglamorous title, "The Market for 'Lemons.'" A lemon, in American slang, is a used car that looks fine but turns out to be terrible. Akerlof's insight, which eventually helped earn him a Nobel Prize, was this: if buyers can't tell good used cars from bad ones, they'll only be willing to pay an average price for any car. That average price is too low for the owners of good cars, so they stop selling. As the good cars leave, the average quality drops, so buyers are willing to pay even less, so more good sellers leave. In the worst case, the market fills up with lemons and collapses. Now replace "used cars" with books, or songs, or articles, or code, or online courses. If readers can't tell the carefully written psychology book from the one produced by a machine in ninety seconds and never read by its "author," they'll become unwilling to pay much for any of them. The honest writer who spent three years on her book can't compete on price with the man who spent three minutes. Discouraged, some honest writers stop. The shelf fills up with lemons. Readers trust the shelf less. And eventually, what gets destroyed isn't just a few jobs. It's the market itself. The shared space where people used to go to find something true. Rabih's room in Dekwaneh is, when you think about it, a perfect little Akerlof machine. Four hundred and sixty phones listening to four hundred and sixty versions of rain on a tin roof, each one shaving a fraction of a cent from the pot that pays every real pianist in the country. No single stream hurts anyone you could point to. That's exactly what makes it work. And that's exactly why Nour's two dollars matter more than they sound. It's never two dollars. It's the slow, invisible draining of a pool that thousands of people drink from. That's the tragedy of the commons, in its modern form. Not a pasture of grass eaten down by too many sheep, but a pasture of trust, grazed until it's bare by people who each take a little, and each tell themselves, like Rabih, that nobody is really hurt. And now for the last suspect in the library. The one I've been avoiding. Us. Because the man with four hundred books only makes money if people buy them. The fake band only makes money if the platform pays for streams. And who pays the platform? Who clicks on the ninety-nine-cent book with the haunted stock-photo cover because it's cheaper than the real one? Who puts on "Rain on Tin Roof, Vol. 17" to fall asleep and never asks who made it? Who shares the AI-generated article without reading past the headline, because it confirmed what we already thought? I'll go first, because this show has a rule. I use AI every day. I'm open about that, and I think I use it responsibly, as a tool that extends work I already know how to do. But I'm not innocent of the rest. I've bought cheap e-books without checking who wrote them. I've let streaming apps play me music without asking whose it was. I've clicked on things I knew, somewhere in my mind, were probably fake, because they were entertaining and it was late. I have been, in small ways, the customer that makes the thief's business model work. Most of us have. The thief needs a buyer, and the buyer is very often us, half asleep, looking for something cheap. The word "author" comes from the Latin auctor. It meant an originator, a founder, someone who makes something grow. And from the same root, through the same history, we get another word: authority. An author was once, quite literally, the person who stood behind a work and vouched for it. The one who could say: this came from me, and I answer for it. That's what a name on a book has always been. Not a brand. A promise. A signature is one human being telling another: I made this, I read it, I believe it's worth your time, and if it isn't, you can blame me. Every honest writer who has ever lived has made that promise with their name. It's the oldest contract in literature, and it's never been written down, because it never needed to be. And that's what the instant author steals. Not a job. Not a sale. He steals the promise. Every time he signs a book he hasn't read, he spends a little of the trust that every honest name has built up, like a counterfeiter spending money he didn't earn. And the trouble with counterfeit money is that it doesn't just cheat the person who takes it. It makes everyone suspicious of every banknote. It makes real money worth less. I know something about that in this country. We lived through it. For years, people here opened their bank statements and saw numbers that said they had savings. The statements were signed. Stamped. Official. And then one day in 2019, the doors closed, and those numbers turned out to mean almost nothing. People's life savings became figures on paper that no one would honor. Rabih, in the story, says nobody went to prison for that, and whatever you think of Rabih, you'll find a great many people here who'd say the same. We learned, in the hardest possible way, what happens to a society when signatures stop meaning anything. It isn't just that money disappears. It's that trust disappears, and trust is much harder to print. So when I watch shelves fill with books whose authors never read them, and playlists fill with songs whose listeners were never born, I don't just see an irritating flood of junk. I see the same thing happening, slowly, to something I care about just as much as money. The meaning of a name. The simple, ancient assumption that when a person says "I wrote this," a person did. Is AI the thief? No. Not in the way we usually mean. The bullet doesn't pull its own trigger. But I'm no longer satisfied with the bullet argument, either, because it lets everyone else walk out of the library too easily. The companies who made volume free and sold it as a shortcut. The platforms that pay for streams without caring who is listening. The instant authors who spend a trust they never earned. And all of us, the customers, who keep buying cheap lemons because they're cheap. Responsibility, I think, isn't a single finger pointing at a single thief. It's more like a web. The person who pulls the trigger bears the most blame. But the one who sold him the gun, and the one who looked away, and the one who bought what he stole, each hold a thread. I don't know how we fix this. Laws will help a little. Detection will help a little. Platforms may, eventually, find ways to tell a human signature from a counterfeit one. But in the end, I suspect, the most important defense is the oldest one. It's the reader who asks, "Who wrote this?" and cares about the answer. It's the listener who looks up the name of the pianist. It's the writer who puts her name on something only after she's read every word. And it's each of us, when we're tempted, like Nour on that curb at midnight, deciding whether our own name is for sale. Outro So, here we are, on a curb after midnight, with a generator roaring, a message unanswered and undeleted, and a statement that says thirty-eight dollars and forty-one cents. Let me leave you with some questions, and I hope you'll argue with them, and with me. If a tool can be used both to help and to harm, is the tool ever responsible? Or does responsibility always belong to the person using it? What about the people who designed the tool, and the companies who profit from it? Is it fair to judge people who come to a craft late, or without any history in it, when they start using AI? Where's the line between a newcomer and a counterfeiter? Is it passion, effort, honesty, or something else? When you buy a cheap book, stream a song, or share an article, how often do you ask who made it? Should you have to? Whose job is it to protect the trust that a name on a piece of work used to carry? The courts have started drawing lines around how AI companies use other people's work. Where do you think those lines should be? Is learning from someone's book different from copying it? Should the authors whose work trained these systems be paid, and if so, how? Think about Nour at the end of the story. Thirty-eight dollars a month, and an offer of forty percent. What would you have done? Be honest. Would your answer change if it were your rent, your child's school fees, your medical bill? And finally: what does your name mean? If you signed it to something tomorrow, what would it promise? Talk about these with your friends, with your family, with the artist in your life who makes thirty-eight dollars a month and keeps going anyway. And come and talk about them with us, too, on our Patreon page at dannyballan.com or on our website, englishpluspodcast.com. Until next time, ask who wrote it. And when it's you, read it first.