Will AGI Make Humans Useless? What the Data Says

6 min read · Last updated: October 25, 2026

Will AGI make humans useless? The data says no, not yet. Stanford payroll research shows employment for workers aged 22 to 25 in AI-exposed jobs is 19% below less-exposed peers, but finds no economy-wide job loss. So far the damage is fewer junior hires, not mass layoffs. That gap is where developers should focus.

Key takeaways

• Stanford finds employment for ages 22 to 25 in AI-exposed jobs is 19% below where it would be if it had kept pace with less-exposed peers.
• The gap comes mainly from reduced hiring, not from layoffs of people already employed.
• The same authors find no evidence of widespread, economy-wide job displacement.
• The World Economic Forum projects 170 million jobs created and 92 million displaced by 2030, a net gain of 78 million.
• Anthropic's CEO warned in May 2025 that AI could eliminate about half of entry-level white-collar jobs within five years.
• Where AI complements workers instead of replacing tasks, employment is flat or rising.

What happened: a Medium post asks if AGI makes humans useless

A post on Medium titled "How A(G)I will make humans useless" has been circulating in developer feeds. Its opening asks you to imagine that a few years from now, your employer tells you your position is no longer needed, because an AI system can do your job faster.

I could not open the full post, because Medium blocked my access, so I am not summarising its argument. What I can do is test its premise against the best data that exists today. Three sources matter: a Stanford payroll study, the World Economic Forum's jobs forecast, and a public warning from the CEO of Anthropic, the AI company behind Claude.

Is AI already replacing workers?

Partly, but in one specific place. Stanford's Digital Economy Lab used ADP payroll data covering millions of US workers through June 2026. It found employment for ages 22 to 25 in AI-exposed jobs is 19% below the trend of less-exposed peers, while experienced workers show no comparable gap.

The authors, Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, are careful about what this proves. They call the results early and descriptive, not causal. They also say the gap has widened steadily since they first documented it in August 2025.

Source	Date	Claim	What it does not say
Stanford Digital Economy Lab	Revised Aug 12, 2026	Ages 22 to 25 in AI-exposed jobs: 19% employment gap	Not a causal estimate; no economy-wide displacement
World Economic Forum	Jan 2025	170M jobs created, 92M displaced by 2030, net +78M	A forecast from employer surveys, not measured outcomes
Anthropic CEO via Axios	May 28, 2025	About half of entry-level white-collar jobs could go within five years; unemployment of 10% to 20%	A prediction by a company that sells AI

Will AGI make humans useless?

No source I opened says that. The strongest claim comes from Anthropic's CEO Dario Amodei, who told Axios in May 2025 that AI could eliminate about half of entry-level white-collar jobs within five years and push unemployment to between 10% and 20%.

He also said "Most of them are unaware that this is about to happen," and that AI producers have "a duty and an obligation to be honest about what is coming." That is a forecast, and it comes from someone who sells AI. Treat it as a warning, not a measurement.

The measured data is milder. Stanford finds no widespread displacement. The World Economic Forum projects a net gain of 78 million jobs by 2030, which equals 170 million created minus 92 million displaced. Both can be true: the total can grow while the entry-level door narrows.

Why does this matter to a developer?

The Stanford mechanism is the part that should worry you. The gap runs mainly through reduced hiring of young workers, not through layoffs. Nobody walks into a meeting and loses a job. A junior role simply never gets posted.

That is invisible, which makes it dangerous. If you already have a job, the data says you are not the group being squeezed. If you are trying to get your first one, you are.

The split inside the data also matters. Declines concentrate where AI mainly substitutes for human tasks. Where AI mainly complements workers, employment is flat or rising, especially for experienced people. The question for your career is which side of that line your daily tasks sit on.

What changes in practice?

Three things change for a working developer.

1. The entry rung gets thinner. Fewer junior roles means more competition for each one. The route in is no longer "learn syntax, apply widely".
2. Judgement becomes the product. Tasks AI substitutes are the repeatable ones. Reviewing, designing and deciding stay with people, which is the complement side of Stanford's split.
3. Compensation holds, headcount moves. Stanford finds the adjustment shows up in employment, not in base pay. Salaries are not the early warning. Job postings are.

I go deeper on each in our posts on whether AI will replace developers, the coding skills that survive AI and the developer job market split.

What should you do this week?

1. Check where your tasks sit. List your last ten tickets. Mark each as one an AI could do alone (substitute) or one it speeds up while you decide (complement). Move your time toward the second list.
2. Practise reviewing AI output. Review is the skill that stays human. Our guide on how to review AI-generated code is a good start.
3. If you are a junior, ship public work. A real project beats a résumé line when fewer roles exist. See how to get a programming job with no experience.
4. Watch postings, not headlines. Track junior openings in your stack monthly. That is the metric Stanford's finding points to.
5. Treat predictions as predictions. A CEO's five-year forecast and a payroll dataset are different kinds of evidence. Weigh them that way.

FAQ

Will AI take my job?

The Stanford data shows risk is concentrated in early-career workers in AI-exposed occupations, with no comparable gap for experienced workers. It finds no evidence of widespread, economy-wide displacement. Your risk depends on whether AI substitutes your tasks or complements them.

What is the difference between AI and AGI?

AGI means artificial general intelligence, a system that could do most intellectual tasks a person can. Today's AI tools are narrower. The Stanford findings above describe the effect of current tools, not AGI, which has not been demonstrated.

Which jobs are most at risk from AI?

Stanford finds declines in occupations where AI mainly substitutes for human tasks, and flat or rising employment where it complements workers. Amodei named entry-level white-collar work as exposed, according to coverage of his Axios interview.

Is the Stanford study proof that AI causes the hiring drop?

No. The authors describe the results as early and descriptive, not causal estimates. They show a widening gap in payroll data through June 2026 and a mechanism of reduced hiring, but not a proven cause.

Pick one task from your own week and decide whether AI replaces it or speeds it up, then browse the free tools on Art of Code to speed up the rest. If you are hiring or job hunting, check /jobs.

Written by Adithya Guttha, Founder of Art of Code.