Definition
People systematically overestimate how much time and effort AI assistance saves them, and the error is specific to AI rather than general optimism. Two named halves: the speedup illusion, where predicted time saving exceeds actual, and the offloading illusion, where predicted effort saving exceeds actual because people overestimate how hard working alone would have been.
What is at stake
Predicted time saving 55.7 seconds, actual 7.5. On easy tasks AI made people ten seconds slower. Unaided estimates were accurate, which puts the miscalibration on AI specifically rather than on optimism in general.
Where the word came from
- The planning fallacyKahneman and Tversky, 1979
- Cognitive offloading as a cost-benefit calculationRisko and Gilbert, 2016
- The efficiency-gain illusionYu et al., 2026, which finds the error absent when the imagined helper is a person
How episcope came to it
Yu's efficiency-gain illusion experiments, 2026. Three pre-registered experiments, 2,691 people. A preprint, and every task ran under five minutes.
What to watch for
This is the measured phenomenon. Cognitive tax is episcope's framing of the same event from the paying side. The framing is not a research finding and does not carry the study's authority.
Papers behind it
Yu, S., Jurafsky, D., Hawkins, R. D., et al. (2026). The efficiency-gain illusion: People underestimate the rate of AI use and overestimate its benefits on simple tasks. arXiv:2605.22687.
Risko, E. F., & Gilbert, S. J. (2016). Cognitive Offloading. Trends in Cognitive Sciences, 20(9), 676-688.