Gen Z and AI: The Heaviest Users Are Turning Against It
Under-30s use AI more than almost anyone and now distrust it more than almost anyone. What divides the generations is not who understands the technology but the first rung of the job ladder.
Executive summary
The generation that uses artificial intelligence most has become the one that resents it most. In the United States, adults under 30 were the least worried age group about AI in 2021; by mid-2026 a majority of them say they are more concerned than excited, and their enthusiasm has fallen further than anyone else's. They have not stopped using it. They use it daily, for school, for search and for comfort, while trusting it less every year.
The split is not about who understands the technology. Teenagers, still years from their first job, are on balance hopeful about what AI will do for them. The mood turns at the point where young people meet the labour market, and that is exactly where the evidence of AI's economic bite is clearest: entry-level hiring in the jobs AI can do has fallen behind. The generation gap over AI is a map of who expects to pay for it.
That makes legitimacy, not capability, the next constraint on AI's spread. The workers and voters of the 2030s are forming their view of AI now, through habit without consent. Whether that hardens into a lasting cohort scepticism or fades as they find their footing is the hinge for the economy, for politics and for the labs racing to win them.
Why age is the right lens on the AI revolution
Every technology wave has a demographic shape, and it usually runs one way: the young adopt first, love it most and drag the old along. The internet, the smartphone and social media all followed that script. AI is breaking it, and the break says something about this revolution that the capability charts cannot. A technology that its most fluent users are turning against is not being pulled into society; it is being pushed.
The best data cuts people by age band rather than by named generation, so this report uses those bands as the proxy: 13 to 17 for today's teenagers, 18 to 29 for older Gen Z and the youngest millennials, 30 to 49 for most millennials and younger Gen X, 50 to 64 for older Gen X and younger boomers, and 65 and over for the rest. Most of the evidence is American, from Pew Research Center and Gallup surveys in 2025 and 2026, and the American picture turns out to be distinctive.
The heaviest users are now the most worried
Start with use. In Pew's February 2026 survey of 5,119 adults, 66% of 18-to-29-year-olds had used an AI chatbot, up from 55% in 2024. Among 30-to-49s it was 61%, among 50-to-64s 42%, and among those 65 and over just 23%. For the first time roughly half of all American adults have tried one. Familiarity follows the same gradient: 66% of under-30s say they have heard a lot about chatbots, against 20% of over-65s, and 31% of under-30s call themselves extremely or very confident using one, against 6% of the oldest group. On one point every generation agrees. Between 61% and 65% of each age band say AI is advancing too quickly, so the disagreement is not about speed. It is about who comes out ahead.
Now set that beside sentiment. In 2021, only 31% of under-30s said the growing use of AI in daily life made them more concerned than excited, the lowest of any age group; 43% of over-65s said the same. In Pew's survey fielded in late June 2026 and published that August, the under-30 figure had risen to 55%, a majority for the first time. Over the same years, the share of under-30s more excited than concerned fell from 25% to 11%. Concern among 50-to-64s went the other way after 2023.
Older Americans settled into wary distance early and stayed there; middle-aged Americans have, if anything, relaxed. The young moved alone, and they moved while using the tools more. Pew's own February survey found that under-30s are the most likely of any group to expect AI to harm society over the next 20 years, roughly half of them, against 14% who expect it to help. Familiarity has not bred affection. It has bred a sharper view of the costs.
Same chatbot, different jobs
Breadth of use and depth of use are not the same thing. On daily use, the peak is not the youngest adults but the working-age core: 34% of 30-to-49s use a chatbot every day, a shade above the 31% of under-30s. After 50 it falls away fast.
The products barely differ by age. ChatGPT is the most-used chatbot in every group, used by 61% of under-30s and 19% of over-65s; Google's Gemini is second everywhere; Anthropic's Claude reaches 10% of under-30s and almost nobody over 65; Character.ai, the role-play companion app, is used by 7% of under-30s. What differs is the job the tool is hired to do.
The young use it as tutor, search engine and confidant. Among teenagers aged 13 to 17, 64% use chatbots and about three in ten do so daily; 54% use them for schoolwork, and 59% say AI cheating happens at least somewhat often at their school. Among under-30 adults, 20% have turned to a chatbot for emotional support or advice, against 2% of over-65s. OpenAI's own study of ChatGPT traffic found that users aged 18 to 25 sent about 46% of the messages from users who stated their age, and that only about 23% of messages from under-26s were work-related; the share rises with age. For the old, AI is mostly something they read rather than talk to: 38% of over-65s say they read the AI summaries at the top of search results, the main way that group meets the technology at all.
That is the hidden half of older adoption. AARP's January 2026 research found that about eight in ten Americans over 50 have used some form of AI, far more than use a chatbot, because AI reaches them inside recommendations, translation, fraud alerts and health apps rather than as a conversation. Most rate themselves beginners: 42% say so, and only 8% say they are advanced or expert. Their stance is wary rather than hostile. Half believe the benefits outweigh the risks, 51% think AI is insecure or unsafe, and 64% worry about it replacing human jobs. Over-65s are also the most likely to say they are simply unsure what AI will do to them, 29% of them, which is the opposite of the young: older Americans lack information, younger ones think they have enough.
The same split runs through creativity. Under-30s who use chatbots are about as likely to say the tools hurt their creativity, 20%, as help it, 25%; among over-65s only 2% say hurt. The heaviest users are the ones who feel the cost.
So the youngest users are not dabbling. They have woven AI into the most personal parts of life, learning and feeling, which is what makes their souring matter.
Anger tracks the first rung of the job ladder
If familiarity drove the gloom, teenagers would be gloomiest. They are not. In Pew's autumn 2025 survey of teens, 36% expected AI to have a positive effect on them personally over the next 20 years and 15% a negative one. Among adults aged 18 to 29, surveyed in early 2026, those numbers reverse: 37% expect harm to themselves and 20% expect good. The mood breaks at the age when people leave school and look for their first job.
The economics give that timing a reason. Stanford's Digital Economy Lab, led by Erik Brynjolfsson, tracks ADP payroll records and finds that by June 2026 employment of 22-to-25-year-olds in the occupations most exposed to AI stood about 19% below where it would have been had it kept pace with their peers in less-exposed jobs, up from 15% a year earlier. The gap shows up in fewer hires rather than more firings, it is concentrated where AI automates tasks rather than assists with them, and experienced workers show no comparable gap. The authors are careful: these are descriptive patterns, not proof that AI caused them, and they find no economy-wide displacement. But the young do not need proof to notice that the first rung is harder to reach.
The mechanism is not mysterious. Junior work is the most codified work: drafting, summarising, first-pass coding, routine analysis, exactly the tasks a model does well. Seniors are protected by what is hard to write down, judgement, relationships and context built over years, which is why the Stanford researchers find that where AI complements workers rather than replacing tasks, employment is flat or rising. The first jobs AI can do are the jobs people used to learn on.
Schools, meanwhile, are teaching the opposite lesson. Gallup found that 74% of American students now attend schools with rules on AI, up from 51% a year earlier, and 48% of students expect to need AI skills for their future careers. A generation is being told to master the tool that is thinning out its first jobs.
They notice. In Harvard's Institute of Politics youth poll of November 2025, 59% of 18-to-29-year-olds called AI a threat to their job prospects, more than named outsourcing to other countries (48%) or immigration (31%), and 44% expected it to take opportunities away against 14% who expected it to create them. In that same Pew survey, published in August 2026, 73% of under-30s expected AI to mean fewer jobs in America over the next 20 years, up from 61% in 2024, the fastest rise of any age group. This report has looked before at why cash transfers would not solve the problem AI creates for work, in our universal basic income episode, number 31; the polling shows the young have reached a similar conclusion on their own.
So the answer to what is really going on is plain. The generation gap is not a knowledge gap. It is an exposure gap: pessimism concentrates on the people whose next decade the technology is most likely to reshape.
Use without love
How do young people cope with a tool they increasingly dislike? Mostly by using it anyway. Gallup's 2026 Voices of Gen Z study, with the Walton Family Foundation and GSV Ventures, surveyed 1,572 Americans aged 14 to 29 and found that use had not moved since 2025: 22% use generative AI daily and 29% weekly. Their feelings moved a great deal.
Even daily users cooled: their excitement fell 18 points. Among the 19% who never use AI, 59% say it makes them angry. Gen Z's trust runs to people, not machines: 69% of employed Gen Z trust work done without AI, 28% trust AI-assisted work and 3% trust work done by AI alone, and the share saying AI's risks at work outweigh its benefits rose from 37% to 48%.
The doubts reach into learning itself, the use the young rely on most. Only 25% of Gen Z say AI helps them think carefully about information, while 42% say it harms that; about eight in ten expect it to make learning harder in future. The fear is less that the machine is wrong than that leaning on it makes them worse.
This is where the scepticism about synthetic content comes from. The young are the generation most fluent in what AI-made media looks like, and the least willing to be fed it. In the 2026 Bentley University-Gallup survey, 66% of 18-to-29-year-olds viewed businesses' use of AI to create advertisements negatively, against 38% of 45-to-59s, and 73% called AI-generated people or voices in ads unacceptable even when disclosed. Older Americans worry about being fooled; younger ones object to being handled. The coping strategy is a split life: AI as a private instrument, used constantly, and a public refusal to accept it as a substitute for a human voice.
What the gap predicts
What the generational split tells us about the future is that use will keep spreading while consent lags behind it, so AI's development will bend toward quieter, more trusted tools, and society will argue about AI less as technology than as a question of work and fairness. Three consequences follow, and they reach well beyond the technology.
For the economy, the risk is a broken pipeline. Firms that stop hiring juniors because AI covers their tasks save money now and lose the people who would have become their seniors. A generation that resents the tool it is told to master is also less likely to build careers on it.
For politics, AI has become a generational issue before it has become a partisan one. Young Americans already rank it above immigration as a threat to their prospects, and in results from that late-June Pew survey published in September 2026 Democrats became more likely than Republicans to be more concerned than excited for the first time. Coalitions are already sorting around AI money, as our episode on AI power in US politics, number 63 traced; a young electorate that blames AI for a hard job market is the natural constituency for whoever runs against it.
For the race between the labs, the lesson is that habit is not loyalty. The companies have won young users' time; they have not won their trust, and young users punish visible AI. Expect the winners to make AI less conspicuous rather than more, embedded in tools people already use instead of announced as a replacement for people, and to compete on trust and safety for younger users as much as on benchmarks. The most intimate uses raise the stakes: when a fifth of under-30s have leaned on a chatbot for emotional support, a lab's product decisions about companionship and teenage users become a question of public trust, not just engagement.
The American pattern may also be a leading indicator rather than a universal one. In Pew's spring 2025 survey of 25 countries, older adults were more likely than 18-to-34s to be mainly concerned about AI in most of them; in Greece the gap was 59% to 18%. The United States, where the labour-market story is loudest, is where the young turned first.
The test is close at hand. Today's teenagers are the most hopeful group in the data. If they sour as they turn 18, the scepticism belongs to the life stage and the job market; if it is already present in them as they age, it belongs to the cohort. These are projections from survey trends, not certainties, and the horizon on which they can be checked is a few years at most. Two things would overturn this reading: a recovery in junior hiring that does not lift young people's mood, which would point to a cultural rather than economic cause, or a rise in teen pessimism before any of them reach the job market.
Bottom line
AI's generation gap runs backwards. The young adopted first and are souring first, and their souring follows the job ladder, not their understanding. For anyone building, regulating or betting on AI, the youngest adults are the early warning system: a technology can win the daily habit of a generation and still lose its consent, and the second loss arrives later and costs more.
Sources
- Pew Research Center: How opinions and use of AI differ by age (June 2026) — chatbot use, daily use, products and uses by age, February 2026.
- Pew Research Center: Young adults in the US are increasingly wary of AI (August 2026) — concern versus excitement by age, 2021 to 2026, and job expectations (survey fielded June 22-28, 2026).
- Pew Research Center: Democrats are now more worried than Republicans about AI and its impact on jobs (September 2026) — concern versus excitement by party, from the same June 2026 survey.
- Pew Research Center: How teens use and view AI (February 2026) — teen chatbot use, schoolwork, cheating and expected personal impact.
- Gallup: Gen Z adoption steady, skepticism climbs (2026) — Voices of Gen Z frequency, emotions and trust in AI-assisted work.
- Gallup: Americans aren't sold on businesses using AI in advertising (2026) — Bentley-Gallup views of AI-made ads by age.
- Stanford Digital Economy Lab: Canaries in the Coal Mine, August 2026 update — employment of 22-to-25-year-olds in AI-exposed occupations.
- Harvard Institute of Politics: Fall 2025 Harvard Youth Poll — young Americans rank AI as a threat to job prospects.
- AARP: Navigating the World of AI (January 2026) — how Americans over 50 use, rate and trust AI.
- NBER: How People Use ChatGPT (Working Paper 34255) — OpenAI usage study, message share by age and work share.
- Pew Research Center: Concern and excitement about AI across 25 countries (October 2025) — older adults more concerned than the young in most countries.
Transcript
Sam: The people who use AI the most are the ones turning against it. And I think I know why.
Alex: You think you know. I think you're wrong, and I can prove it with a calendar.
Sam: A calendar. Of course.
Alex: Give me twenty minutes. Before we start, you have to hear what I saw on Hacker News on Monday. Somebody asked a room of strangers: how do I deal with shipping AI-written code I don't understand?
Sam: That's not a question. That's a cry for help.
Alex: It's what keeps me up. My engineers get first drafts in minutes now, and the job slides from writing code toward auditing it. And auditing is a skill you only get by writing a lot of bad code first.
Sam: So who's going to learn it?
Alex: That's the part I can't stop turning over. The poster asked strangers, which tells you the old answer, turn to the person next to you, has gone quiet.
Sam: Hold that thought. I think it's the same story we're about to tell.
Alex: It might be. I've got a theory.
Sam: You always have a theory.
Alex: Welcome to Dan's AI Intel. I'm Alex.
Sam: I'm Sam. Today's episode is about a generation, and the title says it: the heaviest users are turning against it.
Alex: Here's the setup. Young adults in the United States, the people who grew up on this stuff, used to be the least worried age group about AI. Now most of them say they're more concerned than excited.
Sam: And they haven't stopped using it, which is the part that bugs me. If you hate a thing, you stop doing it.
Alex: Unless you can't. We'll look at who uses what and for which job, where the mood actually breaks, and what a souring generation does to everyone else.
Sam: And we'll fight about why. So let me plant a flag. My starting position is that this isn't mostly about jobs. It's about being handled. I can smell a machine-made brief from across the office, and so can every nineteen-year-old with a phone.
Alex: Mine is that it's a life stage. I've hired a lot of twenty-three-year-olds. Every cohort hits the job market, hates it, blames whatever's new, and then finds its feet. I think the souring is the market, and the market passes.
Sam: Jobs for you, dignity for me.
Alex: Calendar for me. Same mood, same age, different year.
Sam: We'll see. Probably me.
Alex: You'll see. Start with the old story. The internet, the smartphone, social media, they all ran one way. The young adopt first, love it most, and drag everybody else along.
Sam: That's the pattern I know. Teenagers teach their grandparents the video call.
Alex: AI runs it backwards. Back in 2021, Pew found that under-thirties were the least worried group about AI. Just under a third said they were more concerned than excited.
Sam: Pew, the one that cuts by age?
Alex: Pew, yes. And the oldest group was well above that. The old were the nervous ones.
Sam: Fine. That's the intuitive picture. The old are scared and the young are fearless.
Alex: Then this past June Pew asked again, and over half the under-thirties are now more concerned than excited. For the first time, a majority.
Sam: From a third to a majority. That's not drift. That's a swerve.
Alex: And the excited share went the other way. A quarter of them back then, about one in nine now.
Sam: That's not a mood. That's a breakup.
Alex: It's a breakup between people who still text each other every day. Which is the odd bit. Because they use it more than anyone. Pew's February survey of adults found about two thirds of under-thirties have used a chatbot. Among the oldest, fewer than a quarter.
Sam: So the people who know it best like it least.
Alex: And they're confident with it. About a third call themselves very confident using one. Among the oldest, a rounding error.
Sam: Then it's not ignorance. Nobody gets to tell this generation they just don't understand it. My junior copywriters would eat that person alive.
Alex: Right. And every age group agrees on one thing: most of them say AI is moving too fast. So nobody's arguing about speed.
Sam: They're arguing about who comes out ahead.
Alex: Which is where my calendar goes. About half of under-thirties expect AI to hurt society over the next two decades, and a small slice expect it to help. Meanwhile the middle-aged have actually relaxed a little since 2023.
Sam: So everyone isn't getting grumpier. One group moved, alone.
Alex: Alone, and while using it more. Familiarity didn't breed affection. It bred a sharper view of the costs.
Sam: A sharper view of the costs of being handled.
Alex: That's a leap. Park it, we'll test it.
Sam: So who's on it all day? I assumed the kids.
Alex: It's not them. Daily use peaks with people in their thirties and forties, a third of them, a shade ahead of the under-thirties. After fifty it drops fast, and retirees are down in single digits.
Sam: So the daily crowd is the people grinding through a working day. That's a job thing, not a generation thing.
Alex: You're making my argument.
Sam: I'm making a point. Don't get excited.
Alex: Too late. And the products barely differ by age. ChatGPT is the biggest everywhere, Gemini second. Claude reaches a tenth of under-thirties and almost nobody over sixty-five.
Sam: Fine. And Character-something. The role-play one.
Alex: Character.ai. A small slice of the young.
Sam: Because what differs is the job you hire it for, right? My read is that teenagers aren't using it like a work tool.
Alex: They're not. Roughly two thirds of teens use a chatbot, and over half of them use it for schoolwork. Most say cheating with it happens at their school at least somewhat often.
Sam: Which is a strange thing to grow up around. Everyone's using it, everyone's saying everyone's cheating, and you're supposed to know what counts as yours.
Alex: There's a bigger one. A fifth of under-thirties has turned to a chatbot for emotional support or advice. Among the oldest, almost none.
Sam: A fifth. For comfort. Not for a spreadsheet.
Alex: And OpenAI's own study of ChatGPT traffic found something similar. Fewer than a quarter of the messages from under-twenty-sixes were about work.
Sam: So for them it's a tutor and a friend. For my mother it's a paragraph that shows up on its own at the top of a search page. She calls it the robot paragraph.
Alex: That's actually how the oldest group meets it. Plenty of them read the AI summary on search. And AARP found most Americans over fifty have used some AI, far more than use a chatbot, because it arrives inside recommendations and fraud alerts rather than a conversation.
Sam: And they're wary, not hostile. Half of them think the benefits outweigh the risks. Most call themselves beginners.
Alex: And a good few of the oldest say they simply don't know what it'll do to them. The old lack information. The young think they have enough.
Sam: And what they've got enough of is bad news. I'll tell you what that looks like from my desk. My search team watches that robot paragraph every morning, because it decides whether anyone clicks through to us at all. I'd guess a twenty-year-old sees it for what it is faster than I do.
Alex: There's a cost number in there. Under-thirties who use chatbots say it hurts their creativity about as often as it helps. Among the oldest, barely anyone says hurt.
Sam: So the heaviest users are the ones who feel the cost. And that changes how I'd read the usage numbers. Using something every day isn't the same as wanting it, any more than I want my commute. Plenty of habits are just things you can't opt out of.
Alex: They've woven it into learning and feeling, the most personal parts of life. That's why the souring matters.
Sam: Which is exactly why I can't read this as a jobs story only. Nobody sours on their tutor because of a hiring freeze. So give me the calendar. Make it quick.
Alex: Pew asked teenagers last autumn how AI will affect them personally over the next twenty years. More than a third said positive, and a sliver said negative.
Sam: Teens are optimistic. Good for them.
Alex: Then Pew asked under-thirty adults this February. It flips. Now more than a third expect harm to themselves, and only a fifth expect good.
Sam: Different surveys, though. Different kids. And you're asking me to believe the chatbot has nothing to do with a flip that sharp, when the chatbot is the thing they talk to all day.
Alex: Separate surveys, same question, and I'll grant you that. But the break lands right where school ends and the first job search begins. The chatbot didn't change between seventeen and nineteen. The career did.
Sam: That's a coincidence you like a lot.
Alex: It's a coincidence with a mechanism. This reminds me of the dot-com bust, actually, when I was —
Sam: Is this a dot-com story?
Alex: It's a short one.
Sam: It's never short.
Alex: Okay, fine. Mechanism first. Can I show you? Stanford's Digital Economy Lab, Erik Brynjolfsson's group, takes ADP payroll records. And by this June, employment of the youngest workers in the jobs most exposed to AI sits about a fifth below where it should be, compared with peers in less exposed work.
Sam: What counts as exposed, though? Because if the answer is drafting and summarising, I have a whole floor of those people, and I'd like to know before my CFO does. And is that people being fired?
Alex: The gap concentrates where AI automates the tasks rather than just assisting with them. And no, it's not firings, which is the interesting part. It's fewer hires. Experienced workers in the same jobs show no comparable gap. It's the front door, not the exit.
Sam: You are in love with that dataset, aren't you?
Alex: It's payroll records, with experienced workers sitting right beside the young ones as a comparison. It's the most elegant thing I've seen all year. Yes.
Sam: He's blushing, listeners. You can't see it.
Alex: And the authors are careful, which I like. These are descriptive patterns, not proof that AI caused them, and they don't find displacement across the whole economy.
Sam: Good. I don't trust any number that won't say what it can't prove.
Alex: So why juniors? And this is the bit my engineers argue with me about. Junior work is the most codified work. Drafting, summarising, first-pass code, routine analysis. A model is good at exactly that.
Sam: And seniors are safe because?
Alex: Judgement, relationships, context, the stuff that's hard to write down. Where AI complements workers instead of taking over their tasks, employment holds or rises. The first jobs AI can do are the jobs people used to learn on.
Sam: Which is where my copywriters live. I read a number last week that I didn't love. In one survey, nearly half of hiring managers said they'd rather invest in AI tools than hire and train a recent graduate.
Alex: That's the tell. It's a budget line.
Sam: And I looked at my headcount plan after that. A junior leaves. Do I replace her, or buy the tool? The tool never takes a Friday off. It also never grows into my Director of Creative.
Alex: Which is the broken pipeline. You save now, and you lose the people who'd have become your seniors.
Sam: So yes, there's a jobs story. Fine. I'm not stupid.
Alex: I'm writing that down.
Sam: Don't.
Alex: And the feeling of it is specific. My son Diego graduated in economics in May. He's working shifts at a running store while he applies. A survey of recent grads this spring found about two thirds had been ghosted after job interviews.
Sam: Ghosted by a person or a form?
Alex: That's the thing, he can't always tell. He uses ChatGPT on every application, and then he hates the video interviews that reject him with no human in sight. He's cynical about the entire productivity pitch.
Sam: And you sell productivity.
Alex: I do. And I had no good answer.
Sam: Welcome to dinner.
Alex: It's a long dinner.
Sam: Now the part that makes me angry for every sixteen-year-old. Gallup found that three quarters of American students are in schools with rules on AI now. A year earlier it was half.
Alex: And about half of students expect to need AI skills in their careers.
Sam: So the message is: master the tool, and don't get caught. My daughter Maya is applying to college this year. She quizzes herself with it every single day, and she's terrified a detector will flag her essay.
Alex: Are the detectors that bad?
Sam: Vanderbilt turned off Turnitin's AI checker after working out that even a one percent error rate would wrongly flag hundreds of papers a year. And in a Stanford study, detectors flagged most essays by non-native English writers as AI-written.
Alex: So the tool the school hands her is wrong a lot, and the lesson is you're guilty until your prose proves otherwise.
Sam: That's the handled part. It's not the economy. It's a kid being told to use it and punished for using it, and I can't make that a jobs story however hard I squint.
Alex: That's a contradiction. I'll give you that one without a fight. They've noticed the economics too, though. Harvard's Institute of Politics polled young Americans last November. A clear majority of under-thirties called AI a threat to their job prospects, more than outsourcing, more than immigration.
Sam: More than immigration. That's a political fact, not a tech one.
Alex: And this June Pew found that almost three quarters of under-thirties expect AI to mean fewer jobs in America over the next two decades, up from about six in ten in 2024. The fastest rise of any age group.
Sam: Wait, was that Gallup? Or, no, hang on. Gallup's the one about schools.
Alex: It's Pew. You had the two surveys swapped.
Sam: I contain multitudes of confusion.
Alex: You do. And if you want the pipeline problem from the money side, we did universal basic income in episode thirty-one. The polling says young people reached a similar view on their own.
Sam: Cash wouldn't fix a first rung that's missing.
Alex: So the gap isn't knowledge. It's exposure. Pessimism sits with the people whose next decade is on the line.
Sam: Or with the people who feel manipulated. I'm not letting go of that. So how do they cope with a thing they dislike? They just keep using it?
Alex: That's what Gallup's Voices of Gen Z study found, done with the Walton Family Foundation. It surveyed young Americans, teens through twenty-nine, and use hadn't moved since last year. About a fifth daily, nearly a third weekly.
Sam: And the feelings moved.
Alex: A lot. The share who feel excited about AI fell from roughly a third to roughly a fifth. The share who feel angry rose from a fifth to nearly a third.
Sam: So anger's higher than excitement now. That's the line.
Alex: Even among daily users, excitement dropped sharply. And among the people who never use it, most say it makes them angry.
Sam: Angry at something they've never touched. That's not a product review. That's a feeling about what's happening to them.
Alex: Careful. It could still be the jobs.
Sam: It could. But look at what they trust. Of employed Gen Z, about seven in ten trust work done without AI. A bit over a quarter trust AI-assisted work. And almost nobody trusts work done by AI alone.
Alex: They want a human at the end of it. And there's a doubt about learning itself. Only a quarter say AI helps them think carefully about information, and far more say it harms that. Most expect it to make learning harder.
Sam: They're scared of getting worse at thinking, and thinking is what they use it for.
Alex: The fear isn't that it's wrong. It's that leaning on it makes them worse.
Sam: Which brings me to advertising, which is my actual job. Bentley and Gallup asked about businesses using AI to make ads. About two thirds of under-thirties viewed that negatively. Among adults in their middle years, well under half.
Alex: And most of the young called AI-generated people or voices in ads unacceptable, even when it's disclosed.
Sam: Even when it's disclosed. I'd frame that sentence on my wall. Disclosure doesn't fix it. The problem isn't that they were tricked.
Alex: I felt that with Coca-Cola's Christmas ad. The AI-made remake of Holidays Are Coming. People online called it creepy and soulless. I watched it and thought, somebody decided I wouldn't notice.
Sam: Handled.
Alex: Handled. That's your word, and I'm stealing it for this one.
Sam: Coca-Cola's European marketing chief said AI was the best tool for the job, which is what a person says when the question is cost. And the same backlash came back with the next year's ad.
Alex: So they knew.
Sam: Which is why nothing goes out under our name that looks machine-made. I'm not scared of being caught. I'm scared of the look on a twenty-two-year-old's face when they see it.
Alex: Older Americans worry about being fooled. The young object to being handled.
Sam: Thank you. That's what I said an hour ago.
Alex: It was fifteen minutes.
Sam: It felt longer.
Alex: Let's go forward. Three consequences. First, the economy. If firms stop hiring juniors because AI covers their tasks, they save money now and lose their future seniors.
Sam: Okay, but who pays for that? The firm saves the cost today, and the loss lands years from now, on someone else's watch.
Alex: That's the pipeline problem in one sentence. And a generation that resents the tool it's told to master is less likely to build a career on it.
Sam: For me that's a retention problem long before it's a policy problem.
Alex: Same for me. My daughter Lucia is a junior engineer. She's fast with coding agents, and she knows the rung she's on is the one being automated. She defends code review and craft, and we argue.
Sam: Who wins?
Alex: She does, mostly. I told my board the junior pipeline would be fine.
Sam: And?
Alex: Fair, I deserved that. I'm about to say the exact thing I'm warning about.
Sam: Politics next, and this one's the strangest.
Alex: AI became a generational issue before it became a partisan one. Young Americans rank it above immigration as a threat to their prospects. And in the Pew results published in September, Democrats became more likely than Republicans to be more concerned than excited, for the first time.
Sam: First time ever?
Alex: For that question, yes. We covered the money side in episode sixty-three, how AI power picks sides in US politics. A young electorate that blames AI for a hard job market is a natural audience for whoever runs against it.
Sam: Free until the day it isn't. I've seen this pricing movie, and I think the voters have too.
Alex: That's not a data point.
Sam: It's a feeling. I'm allowed one.
Alex: Third, the labs. The companies won young users' time. They haven't won their trust, and young users punish visible AI.
Sam: So the winners hide it.
Alex: They make it quieter. They embed it in tools people already use, rather than announce it as a replacement for people, and compete on trust and safety for young users as much as on benchmarks.
Sam: What about the companion stuff? A fifth of under-thirties has leaned on a chatbot for emotional support.
Alex: Right. And that turns a lab's decisions about companionship and teenage users into a question of public trust, not just engagement.
Sam: Which isn't a technology problem. It's a decision somebody makes in a meeting, probably one with a launch date and no teenagers in the room. I sit in those meetings. The question that gets skipped is always the same: what will the person on the other end feel when they find out?
Alex: One more thing. This may be a leading indicator rather than an American quirk. Pew surveyed twenty-five countries, and in most of them older adults were more worried than the young. In Greece the gap was huge.
Sam: So the American young turned first, because the job story's loudest here.
Alex: That's what the data suggests.
Sam: So settle it. Your calendar against my handling. What decides who's right?
Alex: The hinge. Today's teenagers are the most hopeful group in the data. If they sour as they turn eighteen, it's the life stage and the job market. If they're already souring as they age toward it, it belongs to the cohort.
Sam: And the third door?
Alex: Backlash. Distrust turning into policy, with limits on AI in hiring and media. That's the lower but rising one.
Sam: What would change your mind?
Alex: Two things. Junior hiring recovering without lifting young people's mood, which would say it's cultural, not economic. Or teen pessimism rising before any of them reach the job market.
Sam: See, the second one is mine. If kids are sour at fifteen, jobs can't be the reason.
Alex: Then we'd both learn something. Here's where I am. I still think the sharpest break is the first rung. A first job is the one thing I can actually do something about, so my plan is to protect junior hiring and make review a taught skill, not an accident. But I moved on one thing. Handling matters more than I'd allowed. Disclosure not curing it, that one got me.
Sam: And I moved too, a little. Teenagers being hopeful right up until the job search is a real break, and I can't explain it with ad creative. But I'm not giving up the other half. What I'll watch is whether anything goes out under our name that a twenty-two-year-old would call handled.
Alex: So we split. Jobs and handling, both, in different proportions.
Sam: Which, for a podcast, is a disappointing ending.
Alex: It's the honest one.
Sam: In English, please.
Alex: Fine. The young adopted first and soured first, and their souring follows the job ladder. A technology can win a generation's daily habit and still lose its consent, and the second loss comes later and costs more.
Sam: And that's the line I'll use at my next budget meeting. That's the episode. And if someone in your life keeps asking where AI is heading, sending them this episode is the kindest thing you can do for both of them and the show.
Alex: A note on how this gets made. This show is AI-generated and fact-checked, deeply researched, and worth a second look. Dan builds a custom stack to make it, and he'd rather you check us than trust us.
Sam: And if something worked for you, or dragged, tell us. It shapes what we do next, and we read what comes in.
Alex: Follow the show wherever you listen, on Spotify or Apple Podcasts, so you don't miss the next one.
Sam: Thanks for listening. Go ask someone under thirty what they think. Then really listen.