March 5, 2026
Bot panic vs boss pressure
Labor market impacts of AI: A new measure and early evidence
Study says jobs aren’t vanishing (yet), young hires dip and everyone’s arguing
TLDR: A new measure finds AI’s real-world job automation is limited, but junior hiring in exposed roles may be down. Commenters split between “it’s just faster work” and worries about a quiet squeeze on young workers and doctors using AI—why it matters: it shows where pressure is building.
AI hasn’t torched the job market—yet. A new study rolls out an “observed exposure” score (how much tools like chatbots actually automate real work), finding real-world impact is still a fraction of what’s possible. Jobs with higher exposure are projected to grow less through 2034, and the most exposed tend to be older, female, more educated, and higher‑paid. No surge in unemployment so far—but there’s a spicy wrinkle: hiring of 22–25‑year‑olds into exposed roles appears down about 14%.
Cue the comments. rishabhaiover jokes, “There goes my excuse,” while skeptics like behnamoh insist it’s mostly stealth productivity—use AI, ship more, don’t tell your boss. sp4cec0wb0y says speed is up but management immediately cranked feature demands. yodsanklai raises eyebrows at heavy medical use—“I hope they know what they’re doing”—and nickphx snarks that impact reports prove nothing but self‑importance.
The study’s task‑based approach mixes O*NET job tasks with the Eloundou et al. β score from this paper, emphasizing automation over “assist me” use. The community splits: Team Chill says AI is just a tool making workers faster; Team Squeeze sees a quiet freeze on junior hires and bosses piling on. The meme du jour: AI took my job? Nah, it took my weekend.
Key Points
- •Introduces an “observed exposure” measure combining LLM capability estimates with real-world AI usage, weighting automated, work-related uses more heavily.
- •Finds actual AI use covers only a fraction of theoretically automatable tasks, indicating a gap between capability and adoption.
- •Occupations with higher observed exposure are projected by the BLS to grow less through 2034.
- •Since late 2022, there is no systematic rise in unemployment among highly exposed workers, but there is suggestive evidence of slower hiring of younger workers in exposed occupations.
- •Framework aggregates task-level exposure (from O*NET and Eloundou et al. 2023) to occupations, enabling ongoing monitoring of AI’s labor market effects.