Abstract:A growing literature uses artificial intelligence platform conversation logs to measure occupation exposure. We show that these scores partly measure platform user base rather than the workforce. Holding outcome, sample, controls, and estimator fixed while varying only the platform input changes the post-ChatGPT employment coefficient by a factor of 1.9, and within-vendor consumer-versus-enterprise channels produce estimates that disagree in sign. Reweighting to Bureau of Labor Statistics workforce shares attenuates estimates by 42 to 93 percent. We formalize the non-classical measurement error, derive probability limits and partial-identification bounds for employment elasticities. The bias understates substitution more than augmentation.
| Subjects: | Artificial Intelligence (cs.AI); General Economics (econ.GN) |
| Cite as: | arXiv:2605.21743 [cs.AI] |
| (or arXiv:2605.21743v1 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2605.21743 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Michelle Yin [view email]
[v1]
Wed, 20 May 2026 21:11:48 UTC (2,187 KB)