For all the talk about young people and technology, a surprising number of people in their twenties have barely touched AI. According to Gallup, about half of Americans aged 14 to 29 use it once a month or less, and nearly one in five never use it at all. Plenty have tried it once, to make an image or get past a homework question. Almost everything else they know about it arrives secondhand: headlines, warnings and an online argument that never seems to end.
That isn’t a failing. It’s one of the most ordinary positions in the world right now. If it’s yours, this was written for you, not for the tech crowd.
This isn’t a pitch. Some of what you’ve heard about AI is true, and some of it deserves real anger. What follows is what the research actually shows, what’s fair about the backlash and why the tool is still worth an afternoon of your time.
You’re not imagining the backlash
That same 2026 Gallup survey, the second year it has tracked this, found that excitement about AI fell from 36% to 22% in a single year. Hopefulness fell to 18%. Anger went up, from 22% to 31%.
So if you’re more annoyed than impressed, you’re far from alone. Barely one in five people your age says they’re excited about it.
And the reasons people give aren’t dumb. They’re pretty much the same ones you’d hear in the break room:
“It’s going to take the jobs I’m about to apply for.” This one has real data behind it, and it’s the most important part, so more on it below.
“It makes people lazy.” In that same Gallup survey, eight in ten said AI will likely make it harder for them to learn in the future. A separate Wharton-led survey of 18-to-28-year-olds found 79% think AI makes people lazier. That worry is legitimate. If you let a machine do your thinking, you get worse at thinking. That’s true of calculators, GPS and autocorrect too.
“The internet is full of garbage now.” Fake product reviews, fake photos, comment sections that read like nobody’s home. There’s a name for it: AI slop. In a Gallup survey with Bentley University, two-thirds of 18-to-29-year-olds had a negative view of ads made with AI, and nearly three-quarters said it’s not okay for brands to fake people’s faces or voices.
“It was built on other people’s work, and it burns a lot of power.” Both real concerns. Artists had their work scraped without asking. Data centers use serious amounts of electricity and water. You can care about those things and still decide what to do about the tool itself.
None of that is ignorance. It’s paying attention.
The part about jobs, straight
Here’s the number that should get your attention, because it’s about your age group specifically.
Researchers at Stanford have been tracking payroll data covering millions of American workers. Their latest update, from August 2026, found that people aged 22 to 25 working in jobs where AI can do a lot of the tasks now have employment about 19% below where it would have been otherwise. Older workers in the same kinds of jobs show no such gap.
The researchers are careful to call this an early warning, not proof of cause. But it lines up with what a lot of recent grads are seeing: the entry-level jobs that used to be the first rung of the ladder are getting harder to find.
Now the part that usually gets left out of the headline.
The jobs disappearing are the ones AI just replaces outright. The ones where it helps a person do the work? Those aren’t shrinking the same way. And according to PwC’s 2026 AI Jobs Barometer, which looked at more than a billion job postings, entry-level hiring in the jobs most exposed to AI has flatlined overall. But inside that, one kind of entry-level job is growing fast: the kind that now asks for things like judgment and leadership, stuff that used to be expected only from senior people. In the US data, openings for those roles grew 35% since 2019, while other entry-level roles shrank 10%.
You’ll also see a striking number floating around: job postings that ask for AI skills pay about 62% more. That’s real, but read it carefully. A lot of those are technical jobs. It does not mean “use ChatGPT once and get a 62% raise.” What it does mean is that employers are paying for people who can use these tools well, and that gap is growing, not shrinking.
So the honest version is this. AI isn’t coming for “young people.” It’s coming for tasks. The people who get squeezed are the ones whose whole job is a task a machine can now do. The people who do fine are the ones who can point the machine at the boring part and spend their own time on the part that needs a human.
You don’t need to become a tech person to be the second kind. You need to have actually used the thing.
The weird thing Gallup found
One finding in that survey stands out.
In Gallup’s survey, among people who never use AI, 59% said they feel angry about it. Among people who use it every day, 18% did. Anxiety follows the same pattern: 60% of non-users, 28% of daily users.
That doesn’t prove that using it changes anyone’s mind. It could just as easily be that the people who already liked it were the ones who kept at it. But it does say something: the people most likely to be angry are the ones who’ve only seen AI from the outside, through the worst examples the internet can find.
That’s a bad way to judge anything. Imagine deciding what you think about cooking based only on viral kitchen disasters.
”I don’t really use computers. I play Fortnite.”
That’s a better starting point than it sounds.
You already know how to learn a system by messing with it. Nobody reads the manual for a new game. You drop in, you try stuff, you figure out what works and what gets you killed. That’s exactly how you learn AI. It’s a free-to-play game with no tutorial and no respawn penalty.
It runs on the phone you already have. There’s nothing to install beyond an app, and the free versions of ChatGPT, Claude and Gemini are all fine for starting. You just type, or talk, the way you’d text a friend who happens to know a lot about everything and is sometimes confidently wrong.
And if you like building things in Fortnite’s creative mode, Epic’s tools for making your own islands use a real programming language. AI is very good at explaining code line by line, in plain English, as many times as you need. That’s about as painless as learning to code gets.
Where to actually start
Skip the “make me a funny picture” stuff. You’ve done that. It’s a party trick. Start with something that’s actually bugging you this week. Some ideas that have nothing to do with tech:
- Paperwork you don’t understand. A lease, a pay stub, a letter from your insurance company, a financial aid form. Paste in the confusing part (leave out your account numbers and personal info) and ask, “What does this mean in plain English, and is there anything I should watch out for?”
- A conversation you’re dreading. Asking for a raise, asking for time off, telling a roommate something awkward. Ask it to play the other person so you can practice. Voice mode is great for this.
- Job applications. Paste a job posting and ask what the employer is really looking for. Then have it interview you for the job, and ask it to be tough. Write your own answers. Use it as the coach, not the ghostwriter, because hiring managers can spot a chatbot-written cover letter from across the room.
- Learning a skill. How to change a car battery, how to budget on an hourly wage, how to cook five cheap dinners. Ask follow-up questions. Ask it to quiz you.
- Questions at work. A customer or coworker asks something you don’t know: a product detail, a policy, how some process is supposed to work. AI can get you a solid first answer fast. Then confirm it with someone who knows before you repeat it.
Notice the pattern. You’re not asking it to think for you. You’re asking it to explain, to practice with you, to quiz you. That’s the answer to the “it makes you lazy” worry. The same tool that does your homework can also teach you the material. Which one you get depends on how you use it.
The skill that actually matters for jobs
Let’s be straight about something. Getting comfortable with AI on your own stuff is step one, not the finish line. Remember what the job research said: the entry-level jobs that are growing want judgment. Opening an app doesn’t give you that. Being the person who can tell when the machine is wrong does.
Here’s how to practice that on purpose:
- Test it on something you know cold. Your job, your game, your hometown, your car. Ask it detailed questions and catch where it’s wrong or bluffing. You’ll find mistakes, and learning what they look like is the whole skill.
- Grade its work like a boss would. Have it write a schedule, a plan or a customer email, then pick it apart. What’s missing? What would get you in trouble? What would you change before putting your name on it?
- Argue with it. When it gives you an answer, say “are you sure?” or “what’s the strongest case against that?” Watch how often it changes its mind. That tells you how much to trust it.
That’s the difference between the people the Stanford numbers are worried about and the people the PwC numbers are hiring. It’s not whether you use AI. It’s whether you’re running it or it’s running you.
Two rules that’ll keep you out of trouble
Check anything that matters. AI is often right and sometimes wrong in a very confident voice. For anything involving money, health, the law or your job, treat what it says as a first draft and confirm it somewhere real.
Don’t feed it your private stuff. No passwords, no Social Security number, no bank details, nothing you’d be upset to see leaked. Most free versions can use your conversations to train future models unless you turn that off in settings. Look for it.
The actual point
You don’t have to like AI. You definitely don’t have to like the companies building it, and you shouldn’t stop being loud about the stuff that’s wrong with it. The criticism coming from your generation is some of the sharpest out there.
But hating it from the outside and understanding it are two different things. Only one of them helps when a boss asks what you can do.
Give it one real problem this week. Not a picture. A problem. Then decide what you think.
Correction, September 28, 2026: An earlier version of this article said entry-level jobs heavily touched by AI grew 35% since 2019. PwC’s figure covers a narrower group: entry-level roles in AI-exposed jobs that now ask for senior-level skills such as judgment and leadership. Overall entry-level hiring in those jobs has flatlined. The paragraph has been corrected.
Sources: Gallup / Walton Family Foundation / GSV Ventures, 2026 · Stanford Digital Economy Lab, “Canaries in the Coal Mine,” Aug 2026 · PwC 2026 Global AI Jobs Barometer · Wharton-led survey via Fortune, Jan 2026 · Gallup / Bentley University on AI ads, via Inc., Aug 2026