# Generative AI and the temporary upskilling of knowledge workers > Source: > Published: 2026-11-11 06:19:23+00:00 ## Abstract Upskilling often refers to the process by which workers acquire and expand their skills, enabling them to perform different types of work as market demands change. Here we demonstrate that generative artificial intelligence can act as an ‘exoskeleton’, enhancing workers’ capabilities when attempting new skills, but these gains depend on continued technology use. We run a randomized controlled trial on upskilling by providing management consultants (*N* = 487) access and training in using ChatGPT to solve data science tasks outside their skill sets, which cannot be independently solved by ChatGPT. The treated workers scored higher than the control group on three tasks, although we can reject the hypothesis that they performed as well as expert data scientists. We find no evidence that treated workers are better at answering technical questions without ChatGPT post-experiment, suggesting that their demonstrated technical capabilities do not imply knowledge acquisition. This is a preview of subscription content, [access via your institution](https://wayf.springernature.com?redirect_uri=https%3A%2F%2Fwww.nature.com%2Farticles%2Fs41562-026-02566-6) ## Access options Access Nature and 54 other Nature Portfolio journals Get Nature+, our best-value online-access subscription [Learn more](https://shop.nature.com/products/plus/?region=ROW) Subscribe to this journal Receive 12 digital issues and online access to articles [Learn more](https://www.nature.com/nathumbehav/subscribe) Buy this article - Purchase on SpringerLink - Instant access to the full article PDF. 39,95 € Prices may be subject to local taxes which are calculated during checkout ### Similar content being viewed by others ## Data availability The data that support the findings of this study are available from the corresponding authors (emma.b.wiles@gmail.com and krayer.lisa@bcg.com). Data are available on request for non-commercial, academic review only, provided the requester is a certified academic and is willing to sign a data usage agreement. ## Code availability The code used for analysis is publicly available via GitHub at [https://github.com/emmawiles/Upskill_Public](https://github.com/emmawiles/Upskill_Public). ## References 1. Djankov, S. & Federica, S. The changing nature of work. *J. Int Aff.***72** , 57–74 (2018). 2. Capraro, V. et al. The impact of generative artificial intelligence on socioeconomic inequalities and policy making. *PNAS Nexus***3** , 191 (2024). 3. Dell’Acqua, F. et al. Navigating the jagged technological frontier: field experimental evidence of the effects of artificial intelligence on knowledge worker productivity and quality. *Organ. Sci.***37** , 403–423 (2026). 4. Brynjolfsson, E., Li, D. & Raymond, L. Generative AI at work. *Q. J. Econ.***140** , 889–942 (2025). 5. Autor, D. H., Levy, F. & Murnane, R. J. The skill content of recent technological change: an empirical exploration. *Q. J. Econ.***118** , 1279–1333 (2003). 6. Acemoglu, D. & Autor, D. Skills, tasks and technologies: implications for employment and earnings. In *Handbook of Labor Economics* Vol. 4, Ch. 12, 1043–1171[https://doi.org/10.1016/S0169-7218(11)02410-5]() (Elsevier, 2011). 7. Deming, D. J. & Noray, K. Earnings dynamics, changing job skills, and STEM careers. *Q. J. Econ.***135** , 1965–2005 (2020). 8. Li, L. Reskilling and upskilling the future-ready workforce for industry 4.0 and beyond. *Inf. Syst. Front.***26** , 1697–1712 (2024). 9. What are exoskeletons and how can they help us overcome our human limitations? *Iberdrola*[https://www.iberdrola.com/innovation/what-are-exoskeletons](https://www.iberdrola.com/innovation/what-are-exoskeletons) (2024). 10. Eloundou, T., Manning, S., Mishkin, P. & Rock, D. GPTs are GPTs: labor market impact potential of LLMs. *Science***384** , 1306–1308 (2024). 11. Mollick, E. R. & Mollick, L. New modes of learning enabled by AI chatbots: three methods and assignments. Working paper. *SSRN*[https://doi.org/10.2139/ssrn.4300783](https://doi.org/10.2139/ssrn.4300783) (2022). 12. Agarwal, N., Moehring, A., Rajpurkar, P. & Salz, T. *Combining Human Expertise with Artificial Intelligence: Experimental Evidence from Radiology* (National Bureau of Economic Research, 2023). 13. Wiles, E. & Horton, J. J. More, but worse: the marketplace effects of AI generated job posts. Working paper. *SSRN*[https://doi.org/10.2139/ssrn.5187344](https://doi.org/10.2139/ssrn.5187344) (2025). 14. Dell’Acqua, F. *Falling Asleep at the Wheel: Human/AI Collaboration in a Field Experiment on HR Recruiters**Working Paper* (2022). 15. Carlini, N. A GPT-4 capability forecasting challenge. *Nicholas’s Writing*[https://nicholas.carlini.com/writing/llm-forecast/](https://nicholas.carlini.com/writing/llm-forecast/) (2023). 16. Otis, N. G., Clarke, R., Delecourt, S., Holtz, D. & Koning, R. The uneven impact of generative artificial intelligence on entrepreneurial performance: evidence from a field experiment in Kenya. *Manag. Sci.*[https://doi.org/10.1287/mnsc.2024.06909](https://doi.org/10.1287/mnsc.2024.06909) (2026). 17. Noy, S. & Zhang, W. Experimental evidence on the productivity effects of generative artificial intelligence. *Science***381** , 187–192 (2023). 18. Wiles, E., Munyikwa, Z. & Horton, J. Algorithmic writing assistance on jobseekers’ resumes increases hires. *Manag. Sci.***71** , 10144–10164 (2025). 19. Peng, S., Kalliamvakou, E., Cihon, P. & Demirer, M. The impact of AI on developer productivity: evidence from GitHub Copilot. Preprint at [https://doi.org/10.48550/arXiv.2302.06590](https://doi.org/10.48550/arXiv.2302.06590) (2023). 20. Anderson, J. R. (ed.) *Cognitive Skills and Their Acquisition* (Lawrence Erlbaum Associates, 1981). 21. DeKeyser, R. Skill acquisition theory. In *Theories in Second Language Acquisition* (eds VanPatten B. et al.) 83–104 (Routledge, 2020). 22. Brown, P. C., Roediger, H. L. III & McDaniel, M. A. *Make It Stick: The Science of Successful Learning* (Harvard Univ. Press, 2014). 23. Bjork, E. L. & Bjork, R. A. Making things hard on yourself, but in a good way: creating desirable difficulties to enhance learning. In *Psychology and the Real World: Essays Illustrating Fundamental Contributions to Society* Vol. 2 (eds Gernsbacher, M. A. et al.) 56–64 (Worth, 2011). 24. Agarwal, R. & Prasad, J. A conceptual and operational definition of personal innovativeness in the domain of information technology. *Inf. Syst. Res.***9** , 204–215 (1998). 25. Miron, E., Erez, M. & Naveh, E. Do personal characteristics and cultural values that promote innovation, quality, and efficiency compete or complement each other? *J. Organ. Behav.***25** , 175–199 (2004). 26. Jha, S. & Bhattacharyya, S. S. Learning orientation and performance orientation: scale development and its relationship with performance. *Glob. Bus. Rev.***14** , 43–54 (2013). 27. Lee, D. S. Training, wages, and sample selection: estimating sharp bounds on treatment effects. *Rev. Econ. Stud.***76** , 1071–1102 (2009). ## Acknowledgements We thank BCG and the BCG Henderson Institute for facilitating the experiment and providing access to the participant pool. We are grateful to the BCG data scientists who served as benchmark participants, and to the Economic Impacts Research Team at OpenAI for collaboration on task design. ## Funding This research was supported through the allocation of employee time and internal resources by BCG. No additional external funding was received for this study. Administrative approval was obtained through BCG’s standard internal processes for allocating employee time and company resources. BCG-affiliated authors participated in the conceptualization, design, data collection, analysis, interpretation of results and preparation of the manuscript in their role as co-authors. Decisions regarding manuscript content and submission for publication were made only by the authors. ## Ethics declarations ### Competing interests L.K., M.A., U.A., R.K., D.S. and F.C. are or were employees of BCG. P.M. is an employee of OpenAI. E.W. received consulting fees from BCG through the BCG Henderson Institute during the running of the experiment. The study was conducted using employee time and internal resources provided by BCG. ## Peer review ### Peer review information *Nature Human Behaviour* thanks Solene Delecourt and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. ## Additional information **Publisher’s note** Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. ## Supplementary information ### [Supplementary Information (download PDF )](https://static-content.springer.com/esm/art%3A10.1038%2Fs41562-026-02566-6/MediaObjects/41562_2026_2566_MOESM1_ESM.pdf) Supplementary Sections 1–5, Figs. 1–5 and Tables 1–23. ## Rights and permissions Springer Nature or its licensor (e.g. a society or other partner) holds exclusive rights to this article under a publishing agreement with the author(s) or other rightsholder(s); author self-archiving of the accepted manuscript version of this article is solely governed by the terms of such publishing agreement and applicable law. ## About this article ### Cite this article Wiles, E., Krayer, L., Abbadi, M. *et al.* Generative AI and the temporary upskilling of knowledge workers. *Nat Hum Behav* (2026). https://doi.org/10.1038/s41562-026-02566-6 - Received: - Accepted: - Published: - Version of record: - DOI: https://doi.org/10.1038/s41562-026-02566-6