Between early 2025 and early 2026, the Gallup organization surveyed 1,572 Americans aged 14 to 29 about their relationship with AI. The results were not what most executives would predict. Excitement about AI dropped 14 points, from 36% to 22%. Hopefulness fell 9 points to 18%. Anger rose from 22% to 31%. Nearly half said AI's workplace risks outweigh its benefits, up from 37% the year before.
Then a separate NBC News poll put a sharper number on it: among voters aged 18 to 34, the net favorability rating for AI was minus 44.
The generation that was supposed to be AI's natural constituency has become its most skeptical audience. That has never happened before with a major technology. And the enterprise world is treating it like a communication problem.
The Numbers Don't Lie
Every prior consumer technology follows the same adoption arc. Youth discover it first, embrace it enthusiastically, and adults spend a decade in a moral panic about what it is doing to children. Smartphones, social media, video games: the pattern held consistently. Adults panicked about things kids loved.
AI arrived differently. A Numerator survey of over 5,000 consumers found that 57% of Gen Z non-users were not open to adopting AI at all, compared to 32% of baby boomers. Older Americans were more open to AI than younger ones. The Gallup researcher who ran the youth survey, Zach Hrynowski, noted that even "super users" of AI among Gen Z had grown less excited, less hopeful, and angrier over the year — not just the skeptics.
That last detail matters. The drift toward skepticism is not coming only from people who don't use the technology. It is coming from people who use it constantly and have updated their views based on that experience.
This Is Not Technophobia
The instinct in some quarters is to read this as a generation that doesn't understand what AI will do for them, or hasn't been educated properly about its benefits. That reading misses what the data actually shows.
64% of American teens use AI chatbots. About 30% do so daily. Gen Z weekly AI use sits at 51%. This is not a cohort avoiding the technology. It is a cohort using the technology heavily and forming specific, grounded objections based on that experience.
The Harvard Youth Poll from fall 2025, surveying 2,040 Americans aged 18 to 29, found 59% view AI as a threat to their job prospects — a larger share than those who view immigration (31%) or outsourcing (48%) as threats. That framing is significant. Immigration and outsourcing have been politically charged economic anxieties for decades; AI has eclipsed both as a livelihood concern among young Americans in under three years. The finding held across party lines: 66% of Democrats and 59% of Republicans agreed.
When teens offered open-ended responses to Pew Research in late 2025, they were specific: "It is terrible for the environment and people rely too much on AI." "It's already being used to spread propaganda. It's hard to tell what's real or AI online anymore." These are not panics. They are observations. The kids are reporting what they see, and what they see concerns them.
I've seen a version of this up close. My son and his friends use AI tools. They also push back on them in ways I didn't expect. The environmental cost is a genuine concern, not a talking point, and the numbers back them up: AI data centers consumed roughly 264 billion gallons of water in 2025. But the deeper thing I notice is the uncertainty about what their future looks like inside an economy being reshaped by tools they didn't choose and don't control. That is not technophobia. It is a reasonable response to a real situation.
The Luddites Were Right (The First Time Too)
The Neo-Luddite framing has gone mainstream in 2026. Luddite clubs have spread across US high school and college campuses. A Columbia University conference titled "New Luddism: Technology and Resistance in the Modern Workplace" drew real attendance. Students booed pro-AI commencement speakers on video that went wide. NPR ran a segment on the movement. A wave of mainstream journalism has reconsidered who the original Luddites actually were.
The reconsideration is the important part. Tech journalist Brian Merchant, whose book Blood in the Machine traces the original Luddite movement, makes the case that the historical Luddites were not anti-technology. They were skilled artisans, framework knitters and wool croppers, who responded to an industrial system that was deliberately using new machinery to undercut their craft, destroy their wages, and dismantle the communities built around their labor. They were anti-exploitation, not anti-machines.
The parallel the current generation is drawing is not subtle. As Merchant frames it, today's youth are not rejecting technology per se. They are rejecting the idea that "a handful of powerful companies determine the shape of their future — their working lives, the fabric of their community, their way of life — without a challenge." That is a political position, not a technological misunderstanding. It is grounded in history, and it is spreading.
When Sentiment Becomes Behavior
A generation that is unhappy but still complying is one thing. A generation that is changing its career plans is another.
The Gallup/Lumina Foundation 2026 State of Higher Education Study found that 13% of bachelor's degree students had already changed their major because of AI, and 42% had seriously considered doing so. Computer science and programming enrollment each fell over 10% in 2025-26. Healthcare and engineering rose about 3%. Goldman Sachs reported that workers aged 22 to 25 in the most AI-exposed occupational roles saw a 6% decline in employment from late 2022 through September 2025.
The fear, in other words, has a material basis. These are not abstract anxieties about a speculative future. They are responses to real hiring signals arriving in real time. The Handshake class-of-2025 survey found that computer science majors were the most pessimistic group entering the workforce, with 28% reporting they were "very pessimistic" about their career prospects — up 10 points from the year before.
One in three UK university students, per a King's College London survey conducted in spring 2026, believe AI-driven job losses will be severe enough to cause civil unrest. That is a third of a cohort that will be running organizations, filling technical roles, and voting on AI regulation for the next 40 years.
What Enterprises Are Missing
The enterprise response to this data, where there is one, has been communicative. More messaging about AI opportunity, more "AI won't replace you, it will augment you" positioning, more upskilling announcements. The assumption is that skepticism is a knowledge gap that information can close.
It isn't. The research is clear on this: Gen Z trusts work done without AI more than work done with it (69% to 28%), not because they do not understand what AI can do, but because they have used it, seen its outputs, and formed a judgment. The skepticism comes from using the tools, not from avoiding them. Communicating harder into a position that was formed through direct experience is not a strategy. It is noise.
There are two structural problems underneath it.
The first is talent. Every large-scale AI transformation requires people to run it: to design the workflows, validate the outputs, govern the systems, and catch what the agents get wrong. Those people are being hired from a cohort where 81% expect AI to decrease job opportunities and where the most AI-exposed roles have already shown measurable employment declines. The pipeline organizations are counting on to execute their AI transformation does not trust the transformation it is being asked to execute. That is a change management problem of a different order than most AI strategy documents account for.
The second is the institutional knowledge gap that the throughput trap analysis surfaced earlier this year. If Gen Z skepticism translates into continued drift away from computer science, software engineering, and other highly AI-exposed technical fields, the medium-term supply of people who understand these systems at a deep level shrinks. You cannot govern what you do not understand. You cannot catch errors in AI-generated outputs without the technical foundation to recognize the errors.
The organizations that come through the next five years well are not the ones with the best AI tools or the highest individual productivity metrics. They are the ones that treat the values critique from their incoming workforce as a signal worth taking seriously — not as an obstacle to manage, but as a legitimate set of concerns that should shape how they deploy AI, how they talk about it internally, and what commitments they are willing to make about what AI does and does not mean for the people working inside the organization.
Gen Z is not asking to opt out of the future. They are asking who gets to decide what the future looks like. That is a fair question. It does not have a communication answer. It has a governance answer.