Workforce Rewired Daily Briefing | Tuesday, May 26, 2026
Entry-level jobs in AI-exposed occupations are down 16 percent since 2024. That number did not appear in Sam Altman’s remarks to a Sydney audience this morning, where he said he was “delighted to be wrong” about AI job displacement. It also does not appear in the aggregate BLS data that MIT Technology Review analyzed today, which shows the overall labor market holding steady. Both accounts are accurate. They are describing different people.
By the Numbers
16% decline in entry-level employment in AI-exposed occupations since 2024, per Stanford Digital Economy Lab analysis of ADP payroll data covering millions of workers (MIT Technology Review, May 26)
5.6% unemployment rate for recent college graduates, the highest since the pandemic years, per Federal Reserve Bank of New York (MIT Technology Review, May 26)
1 in 5 companies currently using AI in any business function, per U.S. Census Bureau data, cited as the primary reason broad labor market disruption has not yet arrived (MIT Technology Review, May 26)
OpenAI plans to nearly double its workforce to approximately 8,000 employees by end of 2026, even as its CEO argues the AI jobs apocalypse is unlikely (Euronews / Reuters, May 26)
Layoffs and Company Decisions
Sam Altman Says He Was “Pretty Wrong” About AI Job Displacement
Speaking at a Commonwealth Bank of Australia event in Sydney on Tuesday, OpenAI CEO Sam Altman said the AI boom will not produce the “jobs apocalypse” he previously anticipated. He told CBA chief executive Matt Comyn he was “delighted to be wrong,” adding that he had expected more damage to entry-level white-collar work by now. Altman now attributes the gap to the irreducible human dimension of many jobs, noting that he tested an AI that replied to messages on his behalf and found the experience demonstrated “we really do care about people.” He stopped short of ruling out future disruption, saying his earlier warnings “still may” prove accurate. Despite his reassurances, OpenAI plans to nearly double its own workforce to roughly 8,000 employees by year-end.
Source: Euronews / Reuters, May 26, 2026
Why it matters: The CEO of the company most responsible for accelerating AI deployment into white-collar work says the damage was overstated, at the exact moment that Meta, Cisco, Intuit, and dozens of other firms are citing AI as the rationale for tens of thousands of cuts. HR leaders and workforce planners should pay close attention to both the reassurance and the wave of announced reductions. They are not pointing in the same direction.
Mass Disruption Has Arrived. It Just Started at the Bottom.
MIT Technology Review published a data review today drawing on BLS employment figures, Stanford Digital Economy Lab payroll analysis, the Yale Budget Lab, and a Federal Reserve Board paper on coding employment. At the aggregate level, the conclusion is stable: the BLS unemployment rate in AI-exposed occupations is actually lower than in less-exposed ones, the Yale Budget Lab finds no sign of workers fleeing AI-threatened roles, and only one in five companies is using AI in any business function. Former BLS Commissioner Erika McEntarfer, now at Stanford, puts it plainly: “The data is telling us right now that disruption is not yet here, and we have time to plan.” The entry-level picture is different. Stanford Digital Economy Lab analysis of ADP payroll data found a 16 percent drop in entry-level employment in AI-exposed occupations since 2024, concentrated in roles where AI can complete tasks with minimal human involvement. Older workers in the same occupations saw head counts grow. The Federal Reserve Board found coding employment growth has slowed by 3 percent annually since ChatGPT’s release, even as total coding jobs continue to rise. The earn-while-you-learn model, in which junior hires develop tacit knowledge through practice, is breaking in specific occupations. That pipeline damage will not appear in aggregate data for years.
Source: MIT Technology Review, May 26, 2026
Why it matters: The junior roles that produced every senior software developer, financial analyst, and professional services practitioner are contracting now. Aggregate employment numbers will not show that damage for years. Workforce leaders who rely on macro stability to justify inaction are making a bet on lagging indicators.
What Workforce Leaders Are Watching
Entry-level jobs in AI-exposed occupations are down 16 percent and the earn-while-you-learn model is breaking in software, finance, and professional services. Where does the next generation of senior practitioners come from?
Altman argues that the human dimension of work is what has kept displacement lower than predicted. How should organizations identify which roles carry that irreplaceable human component before executives restructure them away on efficiency grounds?
Only one in five companies is using AI in any business function today. What happens to the organization’s workforce model when that number reaches two or three in five?
Stanford’s Brynjolfsson estimates less than 1 percent of AI investment goes toward understanding the transition. What data is your organization actually collecting about how AI is changing its own workforce, and who owns that question?
This briefing was prepared automatically by the Workforce Rewired research assistant. All stories include direct source links.