Z 世代は AI を否定しているのではない、自身の雇用市場を批判しているのだ（9 分読了）

*Eric Schmidt stood at a lectern at the University of Arizona’s spring commencement and told a stadium full of graduates that the impact of artificial intelligence would be ‘larger, faster, and more consequential’ than anything they had so far lived through.*

The former Google chief executive was attempting, on the published account, to be reassuring. He was speaking about the great human capacity for adaptation. The boos started anyway. They were still going when Schmidt finished.

Twelve days earlier, at the University of Central Florida, the real-estate executive Gloria Caulfield had used the phrase ‘the next industrial revolution’ in her own commencement speech. The students had booed her, too.

The framing the wire reporters reached for, in both cases, was generational confusion. Young people, the framing went, were misreading a technology cycle their elders had already lived through. The framing was wrong.

What the graduating cohort of 2026 was responding to, in both Arizona and Florida, was the most accurate read available on the labour market they were graduating into. They were not booing the technology, they were booing the speech that announced their redundancy.

Consider what the data says about that cohort. Bill McDermott, the chief executive of ServiceNow, told a March conference audience that new-college-graduate unemployment could reach 30% inside two years as AI absorbed the entry-level white-collar workload. The figure was, at the time, treated as deliberately provocative.

Two months later, Goldman Sachs’ April research put the actual number of US jobs being lost to AI at roughly 16,000 per month, with the Gen Z cohort carrying a disproportionate share of the displacement.

The Dallas Federal Reserve, in its own working paper from earlier in 2026, found that the unemployment-rate gap between entry-level workers and experienced workers had widened sharply post-pandemic, specifically in occupations exposed to AI substitution.

Anthropic’s Dario Amodei, the chief executive of the company making one of the most widely-deployed enterprise AI products, has repeatedly forecast that AI will eliminate up to half of all entry-level white-collar jobs. Each of those data points is on the record.

Each is reachable from the same Google search the booing students were running on their phones during commencement.

What makes the cohort distinctive is not that it is sceptical of new technology. Every cohort has been sceptical of the new technology that arrived as it tried to enter the labour market. What is distinctive about the class of 2026 is that it is the first one to enter the labour market while the displacement is being publicly costed by named chief executives in dollar terms, on stage, in named industries, with dated commitments.

Standard Chartered’s chief executive Bill Winters told investors in Hong Kong on Tuesday that the bank would cut more than 15% of its back-office roles by 2030, in HR, risk, and compliance, and that the cuts would ‘replace lower-value human capital’ with AI. Those are the roles new graduates take in their first three years inside a bank.

Meta cut 8,000 jobs on the same week inside a restructuring framed around AI-product reorganisation, with chief executive Mark Zuckerberg framing the trade as converting payroll into AI capital expenditure. The aggregate technology-sector cut so far in 2026, on the running count, is just under 110,000 jobs across 137 companies.

The class of 2026 has been reading these numbers, more or less in real time, since their second year at university.

What gives the booing its accuracy is that the displacement is not, on the available evidence, falling evenly across age cohorts. Earlier generations went through technology cycles in which the workers most exposed to the new technology were also the workers best positioned to acquire the new skills the technology rewarded. The Gen Z exposure pattern is the opposite.

The Dallas Fed paper is precise on this point: the unemployment gap is between entry-level workers and experienced workers, not between technologists and non-technologists. The skill that protects you against this particular wave of automation is not knowing how the technology works. It has ten years of contextual judgment on a workflow that a model can now run in two seconds.

Older workers have that judgment, but younger workers do not. The displacement, in that frame, is being absorbed by precisely the cohort that has the least labour-market power to resist it.

The boardroom response has been to insist that the new equilibrium will produce more interesting work for the cohort that survives the transition. The Multiverse pitch, as one example among many, is that companies will scale the human side of their workforce by training existing employees to operate AI agents rather than by replacing them.

That is a serious thesis. It is also one that depends on companies actually doing the training rather than cutting the headcount. The same StanChart announcement that committed to the 15% back-office cut did not announce a corresponding training programme of equivalent scale.

Meta’s redeployment of 7,000 staff into AI-focused roles, announced the day before its 8,000-person cut, applied to the headcount it was keeping rather than to the headcount it was letting go. The class of 2026 has noticed the asymmetry. The boardroom commentary on the transition has noticed it less consistently.

There is a deeper point about how this cohort encounters the transition. Earlier generations learned about labour-market displacement either in retrospect, through their parents’ experience, or prospectively, through union-organised political education.

Gen Z has learned about it concurrently, through TikTok, LinkedIn, employer-side product launches and CEO-of-the-month commentary, with the dataset arriving faster than any previous generation’s. The result, on the evidence of the booing, is a cohort that has done the arithmetic before the speech-writers have finished writing the speeches.

The MIT researcher Andrew McAfee has been warning publicly that automating entry-level Gen Z jobs will backfire commercially because it eliminates the talent pipeline that produces the experienced workers companies still need. The argument is correct. It is also, in market-implementation terms, taking longer to land than the layoff announcements that ignore it.

The structural fact underneath the cohort’s response is that the AI transition is, on the cleanest read, the first one in modern memory where the productivity gain is being captured in capital rather than redistributed through labour. Microsoft, Alphabet, Meta, Amazon and Apple are committing combined AI infrastructure spending of over $700bn in 2026.

Each of those announcements lands in the same news cycle as the layoff announcement from the same balance sheet. The aggregate spending is rising. The aggregate employment is falling. The trade is being run, in plain accounting terms, by converting wage line items into capex line items.

*The cohort being asked to applaud the transition is the cohort whose entry-level jobs are the wages being converted.*

What the booing is doing, in that frame, is not what the press-section framing said it was. It is not generational confusion about a technology cycle. It is the recognition, by a cohort that has done its own homework, that the speech being delivered to them is the corporate-PR version of the press release that already named them as the line item being eliminated.

The right metaphor for the moment is not the Industrial Revolution. The industrial revolution had a counter-movement, eventually, that organised the workforce into something the technology could not unilaterally absorb.

The class of 2026 has not yet found its equivalent. It has started by booing the commencement speakers. That is, in the strictest possible reading, an unusually accurate first move.