Social posts claim 35% drop in entry-level postings
- X users and Grok amplified claims on May 18-19 that entry-level postings in AI-exposed occupations have fallen sharply since 2023. - The 35% figure matches Revelio Labs research, which said entry-level openings were down by more than 35% versus January 2023. - The next benchmark will be fresh employer and postings data from Strada, the New York Fed and labor-market trackers.
X posts on May 18 and May 19 pushed a claim that entry-level job postings in AI-exposed occupations have dropped about 35% since 2023, with users citing hiring dashboards, recruiter anecdotes and automation concerns. The number did not originate on X. It matches a published estimate from labor-market research firm Revelio Labs, which said in August 2025 that entry-level openings were down by more than 35% from January 2023 levels. The social-media framing went further than the underlying data. Some posts, including one from Grok, tied the decline directly to coding automation and offshoring. Public research reviewed this week shows a more mixed picture: some studies find weaker demand for junior workers in AI-exposed roles, while a May 14 New York Fed analysis said job-posting data showed little sign of a distinct AI-driven drop in labor demand after ChatGPT’s release. (reveliolabs.com) ### Where does the 35% number come from? Revelio Labs published the clearest match for the viral figure. In an Aug. 4, 2025 note, the firm said entry-level openings had fallen by “over 35% fewer job openings than in January 2023,” equal to more than 100,000 fewer new monthly postings. Revelio said the comparison covered entry-level jobs in roles that typically require at least some college. (reveliolabs.com) The World Economic Forum repeated that figure in a March 26, 2026 article, saying postings for U.S. entry-level jobs had “plummeted by 35% in the last 18 months,” citing Revelio Labs. That helps explain why the number is now circulating widely in social posts, newsletters and commentary. (reveliolabs.com) ### Did researchers say AI alone caused the decline? The New York Fed said no. In a May 14, 2026 Liberty Street Economics post, Richard Audoly, Miles Guerin and Giorgio Topa wrote that while overall hiring had slowed since late 2022, job-posting data provided “little indication of a distinct AI-driven decline in labor demand.” Their analysis combined Lightcast posting data with an AI-exposure measure based on Anthropic’s task-level work. (weforum.org) Revelio Labs took a narrower position. In the same August 2025 note, Lisa K. Simon wrote that AI “certainly does not explain the whole decline” in entry-level demand, but said highly AI-exposed entry-level roles had seen the largest declines. Revelio said that, after controlling for industry and time trends, a 10-percentage-point increase in AI exposure was associated with an 11% drop in demand for entry-level roles. (libertystreeteconomics.newyorkfed.org) ### What evidence is there that younger workers are under pressure? The Dallas Fed cited a Stanford-based finding that workers ages 22 to 25 in the most AI-exposed occupations had experienced a 13% decline in employment since 2022. Tyler Atkinson and Shane Yamco wrote on Jan. 6, 2026 that the drop was driven mainly by fewer people moving directly from out of the workforce into jobs, rather than by layoffs. (reveliolabs.com) Strada Education Foundation reported a different signal on May 19, 2026. Surveying nearly 1,500 executives and senior talent leaders, Strada said employers were more likely to expect AI to increase entry-level hiring in 2026 than reduce it, with 27% naming greater AI use as the most significant positive driver among firms expecting higher hiring. (dallasfed.org) ### What should readers be careful about in the X debate? The Stanford graduate-placement anecdote circulating alongside the 35% claim appears unreliable. Searches this week surfaced no official Stanford report supporting the viral “18 of 212” or similar offer counts, while multiple write-ups said those precise figures were fabricated or unsupported by Stanford-published data. (strada.org) The larger point is that the 35% figure is real as a published estimate, but its cause is disputed. Revelio links steeper declines to AI exposure, the Dallas Fed points to weaker employment outcomes for younger workers in highly exposed jobs, and the New York Fed says postings data do not yet show a clear AI-specific collapse in labor demand. The next data points will come from updated postings series, employer surveys and labor-market releases tracking 2026 hiring. (collegeking.com) (reveliolabs.com)