The CEO who spent years warning that AI would eliminate entire job categories now says he got it wrong. OpenAI’s Sam Altman, speaking at a Commonwealth Bank of Australia conference in Sydney this week, walked back his most dire predictions about AI-driven job displacement — and said he’s “delighted to be wrong.”
“I don’t think we’re going to have the kind of jobs apocalypse that some of the companies in our space advocate or talk about,” Altman said. “I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened. I now think I understand more about why it hasn’t, and I’m obviously grateful, but that is an area where my intuitions were just off.”
🔍 THE BOTTOM LINE
The most prominent voice warning about AI job destruction just said the destruction isn’t coming. Whether that’s honest reassessment or convenient rebranding depends on what happens next — and on whether you trust the person selling the product to evaluate its side effects.
What Altman Said Before
This isn’t a small pivot. Altman has been one of the loudest voices in AI warning about job displacement. He said AI would “probably replace most of the jobs people do today.” He told Congress that entire job categories would be “totally, totally gone.” His consistent position was: the displacement is coming, it’ll be painful, and we need to prepare.
Now? The “human part” of employment can’t be replaced, he says. “We really do care about our interactions with people,” Altman told the Sydney audience, adding that this realization had “updated” his thinking on the jobs picture.
What Actually Changed
Altman’s revision comes at an awkward moment for the industry’s growth narrative. Three of the biggest AI companies — OpenAI, Anthropic, and SpaceX — are heading toward public offerings that require convincing investors that AI is both transformative and manageable. OpenAI is targeting $280 billion in revenue by 2030, up from $25 billion today.
Meanwhile, the cost side of the AI equation is starting to look wobbly:
- Uber burned through its 2026 Claude Code budget in four months, with its CTO going viral for the admission
- Nvidia’s VP of applied deep learning, Bryan Catanzaro, told Axios that “for my team, the cost of compute is far beyond the costs of the employees” — meaning AI tools cost more than the humans they’re supposed to replace
- Microsoft has reportedly started canceling Anthropic Claude licenses for its engineers due to cost
And the jobs data? Entry-level white-collar roles have been hit — but not in the straightforward replacement story Altman once told. The displacement is happening, but it’s slower, messier, and more concentrated than the “apocalypse” framing suggested. Our reporting on the entry-level hiring freeze found that companies are freezing junior roles rather than eliminating them outright — a slower bleed, not a sudden flood.
Why the Walkback Matters
There are two ways to read Altman’s reversal:
The generous read: He genuinely got it wrong, and he’s honest enough to say so. The labor market has proven more resilient than expected, humans value human interaction, and the straight-line projection from “AI can do X” to “X jobs disappear” was always too simple. This is a CEO updating his mental model based on new evidence — which is exactly what you’d want.
The cynical read: The CEO of a company seeking $280 billion in revenue by 2030 has discovered that telling your customers their employees are obsolete is bad for business. OpenAI needs enterprises to buy subscriptions, not to retreat in terror. The timing — ahead of an IPO, amid industry cost concerns — is suspicious. And the “I was wrong” framing neatly sidesteps the fact that Altman’s earlier doom-mongering was itself a marketing tool: Fear our product, because it’s that powerful.
The truth is probably both. Altman likely did overestimate the speed of displacement, and he likely also recognizes that the apocalypse narrative has outlived its usefulness as a sales pitch.
The NZ Angle
New Zealand’s early AI adopters are already seeing mixed results, as BusinessDesk reported this month. Air New Zealand was cited as having a permissive AI environment — but even there, the focus is on augmentation, not replacement. The NZ AI Blueprint for Aotearoa, refreshed this month, explicitly avoids the “jobs apocalypse” framing, opting instead for “responsible adoption.”
For a small economy heavily dependent on service industries — tourism, agriculture, education — Altman’s revised view is cautiously reassuring. If AI displaces jobs slowly enough for retraining and adaptation, NZ has time. If he’s wrong again, and the second wave hits faster than the first, the country’s thin employment margins leave little buffer.
❓ Frequently Asked Questions
Q: Is Altman right that the jobs apocalypse won’t happen? It’s too early to call. Entry-level white-collar hiring is clearly under pressure, but total employment hasn’t cratered. The question is whether this is a delay or a cancellation — and nobody, including Altman, actually knows.
Q: What should NZ workers make of this? Cautious optimism, not complacency. Even if AI doesn’t cause mass layoffs, it is changing what skills are valued. The safe bet is still to learn to work with AI, not assume it won’t affect your role.
Q: Why does it matter what one CEO says? Altman’s words move markets and shape policy. His 2025 testimony to Congress influenced regulatory thinking on both sides of the Pacific. When he shifts from “apocalypse” to “actually, we’ll be fine,” that shift ripples outward — potentially slowing the very regulation that might protect workers if his earlier prediction turns out to have just been early.
🔍 THE BOTTOM LINE
The person who did more than anyone to frame AI as a job destroyer now says it probably isn’t. That’s either a welcome correction or a convenient repositioning — and the difference matters. What’s certain is that the labor market is reshaping, just not in the dramatic, sudden way the apocalypse narrative promised. For now, the slow grind continues.
SOURCES
- TIME: “Sam Altman Says AI ‘Jobs Apocalypse’ He Once Predicted Probably Won’t Happen” (May 26, 2026)
- Sydney Morning Herald: “OpenAI boss downplays jobs apocalypse” (May 26, 2026)
- Euronews: “No AI ‘jobs apocalypse’ so far, says OpenAI’s Sam Altman” (May 26, 2026)
- Fortune: OpenAI revenue forecast ($280B by 2030)
- Axios: Nvidia VP on AI compute costs (April 2026)
- AI’s Quiet Crisis: Entry-Level Hiring Freeze
- AI Labor Crisis: 20K Jobs, Meta, Microsoft, $650B
