The Augmentation Trap in Analytics Teams: Short-Run AI Productivity, Skill Erosion, and Decision Resilience
Stout in Business and Management · 2025
DOI: 10.61424/rj1dhn51 ↗Kurz gefasst
Fachartikel in der Zeitschrift Stout in Business and Management, 2025. Verfasst von Md. Nabidul Haque.
Abstract (Originalsprache)
AI can make analytics teams faster, but it can also be faster without human ability. A concept of augmentation trap is developed and tested in this study. The fallacy is that continuously using AI will enhance existing productions but will also decrease the ability to do analysis without AI and make people overly reliant on the AI system. The paper establishes the connections between human-AI complementarity, experiential learning, cognitive offloading, automation bias and organizational resilience. A twelve week-long quasi-experiment is suggested, which is a mixed-methods approach, and with 48 analytics teams and 192 analysts. Teams are divided into groups that have unrestricted AI access, periodic unaided practice and verification with calibrated AI access, delayed or withheld AI access. The design features repeated assessments without AI, an unannounced model-outage exercise, and a task that has a plausible but erroneous AI recommendation. Multilevel growth models test the changes in productivity and unaided skill. Cognitive offloading and team-level models of decision resilience are the focus of mediation tests. The synthetic analysis is only included as a reporting template. It shows an example of a trend where unrestricted access to AI increases throughput, but leads to poorer unaided skill, outage performance and error detection. When calibrated, it will yield smaller gains in speed and higher long run capability. The study adds a time-based perspective to the discussion of AI augmentation, and it reveals why companies should assess performance and capability with and without AI, as well as with incorrect AI.
Zitieren
APA
Haque, M. N. (2025). The Augmentation Trap in Analytics Teams: Short-Run AI Productivity, Skill Erosion, and Decision Resilience. Stout in Business and Management. https://doi.org/10.61424/rj1dhn51
Vancouver
Haque MN. The Augmentation Trap in Analytics Teams: Short-Run AI Productivity, Skill Erosion, and Decision Resilience. Stout in Business and Management. 2025. doi:10.61424/rj1dhn51
BibTeX
@article{Haque2025,
author = {Haque, Md. Nabidul},
title = {The Augmentation Trap in Analytics Teams: Short-Run AI Productivity, Skill Erosion, and Decision Resilience},
journal = {Stout in Business and Management},
year = {2025},
doi = {10.61424/rj1dhn51}
}
EUVIMED