Sam Altman wrong about AI jobs is the headline — but the full story of what he said at the Commonwealth Bank of Australia conference in Sydney on May 26, 2026, and what the data says about jobs and AI in 2026, is more complicated and more consequential than the global headlines suggested. Speaking with CBA CEO Matt Comyn, Altman acknowledged he had been pretty wrong on the technology’s social and economic impact, specifically on entry-level white-collar jobs. He said he was delighted to be wrong about this, and that he thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened. His explanation: the human element of work is stickier than technology might suggest, evidenced by his own experience of having AI respond to his emails on his behalf and then personally returning to answer them himself. The statement made global headlines across Reuters, Fortune, Fortune, Cryptopolitan, TechJournal, and PYMNTS simultaneously. What those headlines did not emphasise, but which the data makes unavoidable: Altman made these remarks precisely eight days after OpenAI filed its confidential S-1 with the SEC, targeting an IPO at a valuation approaching $1 trillion. Dario Amodei of Anthropic followed within days with comparable softened messaging on job displacement. Both IPOs are targeting trillion-dollar valuations. The narrative shift is not subtle.
What Altman Actually Said — and What He Did Not
Altman’s full statement at the CBA conference is more nuanced than the headline suggests, and the nuance matters. He did not say AI will not eliminate jobs. He said the timeline has been slower than he predicted. He distinguished between his predictions on the technology — which he called roughly right — and his predictions on the social and economic impact — which he called pretty wrong. He acknowledged the disclaimer: 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. He added: it still may — a hedge that preserved the long-term risk while walking back the near-term timeline.
The key context that Altman’s remarks did not address: OpenAI’s own policy positions, published in 2026, include a 13-page document calling for taxes on automated labour, a national public wealth fund partly seeded by AI companies, and pilots of a 32-hour working week. These are not the policy proposals of a company that believes AI will have minimal labour market impact. They are proposals designed to mitigate significant displacement. OpenAI’s deployment of the Deployment Company (DeployCo) — a Palantir-style operation that embeds AI engineers inside client organisations to restructure their workflows around AI — is also not the product strategy of a company that believes AI will not change employment patterns. The gap between Altman’s Sydney remarks and OpenAI’s own strategic positioning is the most telling signal in the story.
“I thought there would have been more impact on entry-level white-collar jobs being eliminated by now than has actually happened. I’m delighted to be wrong about this. 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.” — Sam Altman, CEO, OpenAI, Commonwealth Bank of Australia conference, Sydney, May 26, 2026
AI Job Impact Data — What the Numbers Say, May 31, 2026
| Metric | 2026 Data | Source | What It Suggests |
| Tech layoffs (Jan-May 2026) | 113,000+ across 179 companies | TechJournal, Axios | Approaching full-year 2025 total (124,000) in 5 months |
| Layoffs explicitly AI-attributed by companies | 48% of 2026 tech cuts | TechJournal | Nearly half attributed to AI — but may include AI-washing |
| Altman-cited AI-linked cuts (Challenger, Gray & Christmas) | ~50,000 US job cuts | Cryptopolitan / CG&C | Confirms AI attribution but doesn’t prove structural displacement |
| Goldman Sachs estimate | ~16,000 fewer monthly payroll jobs | Cryptopolitan / GS research | AI reducing hiring growth, not causing mass unemployment yet |
| High-AI-exposure occupation unemployment | No significant change | Yale Budget Lab | Structural displacement not yet visible in aggregate data |
| Entry-level hiring in AI-exposed roles | Slowed since late 2022 | Anthropic research | Reduced hiring (not layoffs) is the primary mechanism so far |
| Altman’s previous prediction (June 2025) | Entry-level roles at serious risk within 12 months | Fortune, PYMNTS | 12 months elapsed; serious risk not materialised as predicted |
Why the IPO Timing Is Not Coincidental
The timing of Altman’s and Amodei’s simultaneous walk-back of job displacement predictions is the fact that every financial journalist covering the IPO pipeline noted immediately. OpenAI filed its confidential S-1 on May 22, 2026. Altman spoke at CBA on May 26 — eight days later. Dario Amodei’s comparable softening of his own job displacement warnings followed within days. Both companies are targeting trillion-dollar IPO valuations in the second half of 2026. The single most damaging narrative for any technology company approaching a public listing in a political environment where AI job displacement is a bipartisan political concern is the claim that the company’s products are destroying employment. Walking back that narrative eight days after filing the IPO paperwork is not coincidence. Cryptopolitan put it directly: the timing gives Altman a commercial incentive to soften the narrative around AI-driven job losses at precisely the moment his company seeks public investors.
The irony is that the data does not entirely support either Altman’s original alarm or his May 2026 reassurance. The Yale Budget Lab, which tracks aggregate employment data across AI-exposed occupations, has found no significant changes in occupational mix or unemployment duration in high-AI-exposure jobs since ChatGPT launched in late 2022. This is the finding that Altman’s Sydney remarks drew on. But the mechanism of impact that economists increasingly identify is not mass layoffs — it is reduced hiring. Anthropic’s own research found suggestive evidence that hiring of younger workers has slowed in AI-exposed occupations since late 2022. Goldman Sachs research found AI reduced monthly payroll growth by approximately 16,000 jobs. These are not the numbers of a technology with no impact — they are the numbers of a technology whose impact is real and growing but operating through the less visible mechanism of slower hiring rather than the more visible mechanism of mass terminations.
“Altman made these remarks weeks before OpenAI expects to confidentially file for a US IPO targeting a $1 trillion valuation and raising at least $100 billion. The timing gives Altman a commercial incentive to soften the narrative around AI-driven job losses at precisely the moment his company seeks public investors.” — Cryptopolitan, analysis of Sam Altman’s CBA conference remarks, May 2026
Dario Amodei’s Parallel Walk-Back — A Different Message
Dario Amodei’s own position shift in May 2026 is meaningfully different from Altman’s in a way that the combined headline coverage has obscured. Altman walked back the timeline and severity of his job displacement predictions — reducing the urgency of the near-term risk. Amodei maintained that up to half of entry-level white-collar jobs will dissolve within five years and that unemployment could reach 10 to 20 percent, while simultaneously calling on AI producers to be honest about what is coming. This is not a walk-back — it is a restatement of a severe prediction, delivered with more emphasis on the timeline being five years rather than imminent. The two most prominent AI CEOs are making diametrically opposite public predictions about AI’s employment impact in the same week, which is itself a story: if the companies with the best operational visibility into frontier AI capability cannot agree on whether the labour market impact is minimal or catastrophic, the public debate is operating without a reliable signal from the people who should know best.
AI Magazine, which tracked both executives’ statements across the week of May 26, characterised the pattern as a major corporate pivot, describing both executives as backtracking on their claims and aligning with traditional financial leaders ahead of looming IPOs. Whether that characterisation is accurate depends on which statement you treat as the prior belief and which as the revision. Altman’s June 2025 Uncapped podcast warning — that a lot of jobs will go away — and his May 2026 Sydney conference reassurance — that he is delighted to be wrong — are not obviously consistent with each other. The IPO filing sits precisely between them.
“Both OpenAI and Anthropic are reportedly preparing to launch their respective IPOs this year, each company with an estimated valuation of $1 trillion. For the OpenAI CEO, his comments walk back his prophecy on AI’s impact on labour.” — Fortune, analysis of Sam Altman and Dario Amodei job prediction reversals, May 2026
Key Takeaways
• OpenAI CEO Sam Altman told the Commonwealth Bank of Australia conference in Sydney on May 26, 2026 that he was ‘pretty wrong’ about AI’s economic impact — specifically that he expected more entry-level white-collar job elimination by now than has occurred. He called his technology predictions ‘roughly right’ and social/economic predictions ‘pretty wrong.’
• The statement arrived eight days after OpenAI filed a confidential S-1 with the SEC targeting a $1 trillion IPO valuation — creating an unavoidable commercial context for a narrative that shifts attention from AI-driven job displacement precisely as OpenAI seeks public investors.
• The data tells a more complex story: 113,000 US tech layoffs in 2026 (48% explicitly AI-attributed by companies), Goldman Sachs estimates of 16,000 fewer monthly payroll jobs due to AI, and Anthropic research finding slowed hiring of younger workers in AI-exposed occupations since late 2022 — displacement through reduced hiring, not mass terminations.
• Dario Amodei maintained his severe prediction — up to half of entry-level white-collar jobs dissolving within five years, 10-20% unemployment possible — in direct contradiction to Altman’s softened messaging during the same week. The two CEOs with the most operational visibility into frontier AI cannot agree on the labour market trajectory.
• OpenAI’s own 2026 policy document calls for taxes on automated labour, a national public wealth fund seeded by AI companies, and 32-hour workweek pilots — positions inconsistent with a company that believes AI will have minimal labour market impact.
• Yale Budget Lab found no significant aggregate changes in AI-exposed occupational unemployment — consistent with Altman’s reassurance on timing, but the mechanism economists identify (reduced hiring, not mass layoffs) is less visible in aggregate data and more damaging to long-term employment entry for younger workers.
Conclusion
Sam Altman’s Sydney walk-back will be studied as one of the more transparent examples of CEO messaging shifting with corporate incentive in the AI era. The original predictions — entry-level roles at serious risk within twelve months — were made in June 2025 without an IPO on the near horizon. The revised messaging — pretty wrong, delighted, the human part is stickier than technology suggests — was delivered eight days after confidential IPO filing with a $1 trillion valuation at stake. Both things can be simultaneously true: the timeline was genuinely slower than Altman predicted, and the timing of his public acknowledgement was not coincidental. The data neither fully vindicates nor fully refutes his original alarm. The labour market impact of frontier AI in 2026 is real, operating primarily through reduced hiring rather than mass terminations, concentrated in entry-level and early-career roles, and still in its early stages. That is not the jobs apocalypse Altman warned about. It is also not the reassuring picture his Sydney remarks implied.
Frequently Asked Questions
What did Sam Altman say about AI and jobs in May 2026?
At a Commonwealth Bank of Australia conference in Sydney on May 26, 2026, Altman said he was ‘pretty wrong’ about AI’s economic impact — specifically that he expected more entry-level white-collar job elimination than has actually occurred. He said he was ‘delighted’ to be wrong and explained that the human element of work is stickier than he anticipated. He maintained that his technology predictions were ‘roughly right’ but social and economic impact predictions were ‘pretty wrong.’
Is Sam Altman walking back his AI job predictions because of OpenAI’s IPO?
The timing is notable. Altman made his job prediction reversal eight days after OpenAI filed a confidential S-1 IPO registration on May 22, 2026, targeting a $1 trillion valuation. Multiple publications including Cryptopolitan and Fortune noted that the IPO context gives Altman a commercial incentive to soften job displacement narratives at precisely the moment OpenAI seeks public investors. Altman himself acknowledged the criticism of fear-mongering but maintained the risk ‘still may’ materialise.
What does the data say about AI and job losses in 2026?
The picture is mixed. The Yale Budget Lab found no significant changes in aggregate unemployment in AI-exposed occupations. Goldman Sachs estimated AI reduced monthly payroll growth by approximately 16,000 jobs. Challenger, Gray and Christmas counted nearly 50,000 US job cuts explicitly attributed to AI in 2026. Anthropic’s own research found slowed hiring of younger workers in AI-exposed roles since late 2022. The mechanism is primarily reduced hiring rather than mass layoffs — which is less visible in aggregate data but more damaging to career entry for younger workers.
What did Dario Amodei say about AI jobs?
Unlike Altman, Amodei maintained his severe prediction in May 2026: up to half of entry-level white-collar jobs could dissolve within five years, with unemployment potentially reaching 10-20%. He called on AI producers to be honest about what is coming. Amodei’s position is the direct opposite of Altman’s May 2026 messaging, creating a situation where the two most prominent AI CEOs hold diametrically opposite public views on the labour market impact of frontier AI.
Does OpenAI think AI will affect jobs?
OpenAI’s own 2026 policy documents call for taxes on automated labour, a national public wealth fund seeded by AI companies, and pilots of a 32-hour working week — positions that are inconsistent with a company that believes AI will have minimal labour market impact. This creates tension with Altman’s Sydney reassurance: the company’s policy advocacy and its CEO’s public statements are pointing in opposite directions on the question of how serious the employment disruption will be.
References
Fortune. (2026, May 26). Sam Altman and Dario Amodei are both walking back AI jobs apocalypse predictions as they eye IPOs. https://fortune.com/2026/05/26/sam-altman-dario-amodei-walking-back-ai-jobs-apocalypse-prophecies-ipo/
TechJournal. (2026, May 29). Sam Altman says he was wrong about AI jobs — the data disagrees. https://techjournal.org/sam-altman-wrong-ai-jobs-apocalypse
Cryptopolitan. (2026, May 27). Sam Altman calls AI job fears overstated, but labor data tells a different story. https://www.cryptopolitan.com/sam-altman-calls-ai-job-fears-overstated/
PYMNTS. (2026, May 26). OpenAI’s Sam Altman retracts AI job cut prediction. https://www.pymnts.com/artificial-intelligence-2/2026/openais-sam-altman-retracts-ai-job-cut-prediction/
AI Magazine. (2026, May 27). Why are AI’s top CEOs pedalling back on job predictions? https://aimagazine.com/news/altman-amodei-drop-ai-job-predictions
Breitbart / Fortune wire. (2026, May 27). OpenAI’s Sam Altman and Anthropic’s Dario Amodei reverse predictions on AI job displacement as IPOs loom. https://www.breitbart.com/politics/2026/05/27/openais-sam-altman-and-anthropics-dario-amodei-reverse-predictions-on-ai-job-displacement-as-ipos-loom/
4Corner Resources. (2026, May 28). Sam Altman admits he was wrong about AI job losses. https://www.4cornerresources.com/job-market-news/sam-altman-ai-jobs-may-2026/
