AI and Employment: So Far, So Good - Marginal REVOLUTION

Open original

Census Bureau surveys tracking business AI use from 2023 to 2026 show that even as AI adoption tripled, firms report almost no net employment changes. About 95% of businesses say AI hasn't affected their headcount either way.

  • AI adoption jumped from 3.7% of firms in September 2023 to 10% by late 2025 (18% when counting any AI use), but employment impact remained flat: roughly 2-3% reported job increases, 2% reported decreases, and 95%+ reported no change across both survey periods.
  • Among firms actually using AI, 44% say it supplements existing work, 10% say it replaced employee tasks, and 11% say it created new tasks—but most firms (64%) made no business changes to implement AI and only 1% hired new AI-skilled workers.
  • Task substitution is growing within the small subset of firms where it's happening: the share reporting AI took over "a large number" of tasks jumped from 2.4% to 7.1%, but this group still represents only about 2% of all firms.

The Census Bureau has been tracking AI adoption and its employment effects since 2023, and the data shows remarkably little disruption so far. AI use in producing goods and services jumped from 3.7% of businesses in September 2023 to about 10% by late 2025 (18% when measured more broadly). When asked directly whether AI affected total employment over the previous six months, the results stayed nearly flat across both surveys: around 2.8% reported increases, 2.6% reported decreases, and 94.6% saw no change. Two years later, those numbers barely budged—2.3% up, 2.0% down, 95.7% unchanged. The Information sector showed more movement than others, but the pattern held across different firm sizes. Among firms actually using AI, the applications remain narrow and mostly supplementary. Forty-four percent say AI supplements work employees already do; only 10% say it replaced a task someone used to handle. The most common uses are writing or editing documents (85% of generative AI adopters), followed by information searches (50%) and document summarization (45%). Most firms made no operational changes to adopt AI—64% changed nothing about their business, while just 15% trained existing staff and another 15% built new workflows. Only about 1% hired anyone specifically for AI skills. There's one wrinkle: among the small subset of firms where AI actually did take over employee tasks, substitution is accelerating. The share reporting AI took over "a large number" of tasks grew from 2.4% to 7.1%, and those seeing "a moderate number" rose from 13% to 22%. But this group remains tiny—roughly a tenth of AI adopters, who themselves represent about a fifth of all firms. The author emphasizes this comes from direct Census data on a massive sample, making it unusually reliable evidence that most businesses using AI haven't yet felt employment effects.

Ask about this article