Picture a university seminar room, sometime in the last year. A professor has just finished explaining how a large language model works — attention mechanisms, token prediction, the usual — when a student near the back raises her hand. She doesn’t ask how it works. She asks: “Who decided this should exist?”
That question hangs in the air a lot these days. And it’s coming, overwhelmingly, from young people — the generation that grew up with algorithmic feeds curating their reality, with recommendation engines nudging their politics, with facial recognition at airports and biometric data harvested through apps they downloaded at age twelve. They are not naive about technology. If anything, they are its most experienced subjects.
What This Tells Us
The dominant narrative around young people and AI is flatly wrong. The conventional wisdom — repeated in boardrooms, policy papers, and breathless tech coverage — is that Gen Z and younger millennials are AI’s most natural adopters. They’re digital natives, the story goes; of course they’ll embrace it. This framing is not just lazy, it’s actively misleading. It conflates usage with endorsement. It mistakes familiarity for trust.
I believe we are misreading one of the most important signals available to us. When young people raise concerns about AI — about bias, about job displacement, about unchecked power concentrating in a handful of private companies — they are not being technophobic. They are being precise. They have lived inside algorithmically mediated systems long enough to understand, viscerally, what goes wrong when those systems are deployed at scale without accountability. We should be treating their alarm as expert testimony, not as youthful anxiety to be soothed.
The Assumed Story: Digital Natives Love AI
The “digital native” framing was always a stretch, but it has calcified into received wisdom. The logic runs: young people grew up with smartphones, so they’re comfortable with technology, so they’ll be comfortable with AI. Each step in that chain is weaker than the last.
Comfort with a tool is not the same as comfort with being governed by it. Young people use AI tools — of course they do, because those tools are now embedded in the search engines, content platforms, and productivity software they have no practical choice but to use. But using a system under structural pressure is categorically different from endorsing it. We wouldn’t say that factory workers in 1910 “embraced” dangerous machinery simply because they showed up to operate it.
The data picture here is more complicated than the boosterish narrative suggests. Survey after survey of younger cohorts in recent years has found high levels of concern about AI’s social effects — on employment, on privacy, on fairness — sitting alongside high levels of daily use. That’s not contradiction. That’s what it looks like when a generation is sophisticated enough to separate a tool’s utility from its systemic risks. As Blockgeni has reported, the AI systems people use every day carry embedded biases that their users often sense before researchers formally document them. Young, frequent users are often the first to notice the distortions.
The Overlooked Angle: Experience as Expertise
Here is the thing that keeps getting missed: the young people raising alarms about AI are not doing so from ignorance. They are doing so from exposure. They grew up inside the first generation of large-scale algorithmic systems — social media recommendation engines, content moderation tools, predictive policing software used in their schools and neighbourhoods. They have watched those systems fail, in documented, personal ways.
That lived experience constitutes a form of expertise that the industry consistently undervalues because it doesn’t come with a PhD or a stock option. When a twenty-two-year-old tells you that AI-generated content is eroding her ability to distinguish what’s real, she is not being dramatic. She is reporting a perceptual reality that researchers are now scrambling to measure. When a first-generation university student says he’s terrified that AI will displace the kind of entry-level professional job he’s training for, he is identifying a structural economic risk that labour economists increasingly confirm is concentrated precisely in the roles that credential-earners historically relied upon to build careers.
There is an underreported tension here worth naming explicitly: the same AI capabilities being marketed as productivity tools for established professionals are most disruptive at the entry level — the very rung that young people are trying to step onto. A senior lawyer using AI to draft briefs faster faces a competitive enhancement; a paralegal graduate faces potential structural unemployment. The technology looks entirely different depending on where you stand on the career ladder, and young people are standing at the bottom of it. Their alarm about AI and jobs is not abstract. It is a forecast about their own futures, informed by proximate observation of the industries they are trying to enter.
This connects to a broader dynamic that should concern anyone thinking seriously about AI governance. The people with the most granular, lived experience of algorithmic systems at scale are systematically excluded from the institutions that govern those systems. As credentialing systems themselves come under pressure from AI, the disconnect between the governed and the governors is likely to widen, not narrow.
What This Changes: The Stakes of Getting It Wrong
If we continue to misread young people’s AI scepticism as enthusiasm, we will make a series of compounding policy and product errors. We’ll design systems without adequate input from the communities most affected. We’ll deploy AI in education, hiring, and public services without the safeguards that the most experienced users would demand. And we’ll lose the window — it’s already narrow — to build the kind of public legitimacy that makes AI governance actually work.
There’s a security dimension here too. Young people aren’t just concerned about economic disruption; they’re concerned about who controls the systems shaping their information environment. Those concerns track closely with the warnings coming from intelligence and security communities. Five Eyes agencies have warned that AI-enabled threats are arriving faster than institutions can adapt, and the manipulation of information environments is high on that list. A generation that has lived through algorithmic radicalisation, coordinated inauthentic behaviour, and synthetic media knows in their bones what those warnings mean.
The failure to take young people’s concerns seriously is also, frankly, a research failure. The AI field has a well-documented tendency to treat bias and fairness concerns as secondary to capability benchmarks. That ordering of priorities reflects who is in the room when priorities are set — and young people, particularly those from communities historically harmed by automated systems, are rarely in that room.
Where I Could Be Wrong
The strongest counterargument to my position is a version of the age-old optimism about technology adoption: that young people’s concerns, while genuine, will moderate as AI systems improve. The steel-manned version of this objection holds that earlier generations were similarly alarmed by the internet and social media — not entirely without reason — but that the net effects of those technologies, while mixed, did not produce the catastrophic outcomes that critics predicted. Regulators found (partial) footing. Social norms adapted. The generation that grew up with smartphones figured out how to navigate them, and so will this one figure out how to navigate AI.
There is real weight to this. Generational technology anxiety is not a new phenomenon, and some of the alarm around AI is undeniably shaped by novelty and uncertainty rather than demonstrated harm. It would be intellectually dishonest to ignore this entirely.
But here is why I think this analogy fails: the pace and scope of AI deployment is categorically different from prior technology waves. Social media was disruptive; AI is infrastructural. We are not talking about a new category of app — we are talking about systems that are being embedded into hiring pipelines, credit decisions, medical diagnostics, and military targeting simultaneously, at speed, with governance frameworks that lag by years. The investment wave currently flooding the sector is accelerating deployment faster than any prior technology. The “wait and see” approach that worked, imperfectly, for social media is not viable when the systems in question are making consequential decisions about people’s lives right now. Young people understand this urgency in a way that those further from the deployment frontier often don’t. The legal system is only beginning to catch up, and it is doing so slowly.
What I Expect Next
My prediction is this: within the next three to five years, young people’s concerns about AI will produce a formal political constituency — not a diffuse cultural mood, but an organised force capable of influencing AI regulation in democratic systems. We’ve seen precursors in student-led movements around data privacy and algorithmic accountability. The generational cohort entering voting age now is the first one for whom AI is not a future technology but a present condition. Their political mobilisation around AI governance is not a matter of if, but when.
The falsifying signal would be evidence that younger cohorts are, in aggregate, shifting toward positive endorsement of AI deployment without accountability constraints — that usage is in fact translating into trust as the technology matures. If large-scale longitudinal surveys consistently show that AI trust among under-thirties is rising alongside deployment rates, I would need to substantially revise this argument. But until that evidence exists, the responsible analytical position is to treat the alarm as the signal — not the noise.
