Position: Adopting AI in Practice Does Not Guarantee the Productivity Boost
Pith reviewed 2026-06-30 11:57 UTC · model grok-4.3
The pith
Adopting AI in organizations does not guarantee productivity gains because human and environmental factors can substantially reduce or negate the benefits.
A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.
Core claim
The paper claims that regardless of apparent performance advances in AI technology, human and environmental factors of the organization may substantially attenuate or even negate the effective productivity benefits. It identifies five key moderating factors: human resource composition, baseline capability of individuals, learning curve of practitioners, incentives for fair use, and flexibility of objectives. Drawing on and revising the partial equilibrium model of Gries and Naudé (2022), the authors redefine effective organizational determinants and discuss practical implications for industry and education.
What carries the argument
Five moderating factors—human resource composition, baseline capability of individuals, learning curve of practitioners, incentives for fair use, and flexibility of objectives—that determine whether AI deployment produces realized productivity improvements.
If this is right
- Organizations should assess human resource composition and baseline capabilities before deploying AI tools.
- Training and education efforts must explicitly address practitioner learning curves to realize AI benefits.
- Incentive systems should be redesigned to encourage fair and effective AI use rather than misuse.
- Organizational objectives need to stay flexible enough to accommodate AI integration without rigid targets.
- Stakeholders in industry and education are called to revise practices based on the updated model.
Where Pith is reading between the lines
- Productivity studies on AI should control for these organizational factors instead of crediting gains solely to the technology.
- The revised model could be tested in specific sectors like software development or customer service to quantify the moderation effects.
- Policy guidelines for AI adoption might incorporate checks for these five factors to avoid inefficient investments.
Load-bearing premise
The partial equilibrium model of Gries and Naudé (2022) can be revised to center these five factors as the primary determinants without introducing inconsistencies or requiring additional empirical grounding.
What would settle it
A study measuring productivity changes in matched organizations using the same AI tools but differing in one moderating factor, such as incentive structures for fair use, that finds no attenuation or negation of gains would falsify the central claim.
read the original abstract
This position paper argues that adopting AI in organizational practice does not guarantee productivity gains, because human and environmental factors critically moderate the relationship between AI deployment and realized productivity improvements. Following the advent of high-performance generative models, AI use has been rapidly encouraged in some sectors while being restricted in others. Most practitioners assume that AI brings productivity boosts owing to enhanced technical capabilities, but regardless of apparent performance advances in AI technology, human and environmental factors of the organization may substantially attenuate -- or even negate -- the effective productivity benefits. We identify five key moderating factors: human resource composition, baseline capability of individuals, learning curve of practitioners, incentives for fair use, and flexibility of objectives. Drawing on the partial equilibrium model of Gries and Naud\'e (2022), we argue that existing economic frameworks may inadvertently overlook these factors. We revise the existing framework to redefine effective organizational determinants and shed light on practical implications including industry and education, responding to alternative views and calling for action of stakeholders.
Editorial analysis
A structured set of objections, weighed in public.
Referee Report
Summary. This position paper claims that adopting AI in organizational practice does not guarantee productivity boosts, as human and environmental factors can substantially attenuate or even negate the benefits. The authors identify five key moderating factors: human resource composition, baseline capability of individuals, learning curve of practitioners, incentives for fair use, and flexibility of objectives. Drawing on the partial equilibrium model of Gries and Naudé (2022), they argue that existing economic frameworks overlook these factors, revise the framework to redefine effective organizational determinants, and discuss practical implications for industry and education.
Significance. If the central argument holds, the paper would provide a valuable cautionary perspective on AI adoption, emphasizing that technological advances alone are insufficient without addressing organizational moderators. This could inform stakeholders in industry and education. However, the significance is limited by the lack of new empirical evidence or explicit model derivations to support the revision and the selection of the five factors.
major comments (2)
- [Abstract] Abstract: The revision of the Gries and Naudé (2022) model is stated as redefining effective organizational determinants, but no explicit revised equations or equilibrium conditions are presented to show how the five factors are incorporated as primary determinants or to verify internal consistency of the modification.
- [Discussion of the five factors] Discussion of the five factors: The identification of the five moderating factors (human resource composition, baseline capability, learning curve, incentives for fair use, flexibility of objectives) is presented as key to attenuating productivity gains, but without derivation from the base model or any validation steps, it is unclear whether these are load-bearing primary determinants.
minor comments (2)
- [Abstract] The abstract is lengthy and repetitive in places; condensing it would improve clarity without losing the core position.
- [References] Ensure consistent citation formatting for Gries and Naudé (2022) and any additional references throughout the manuscript.
Simulated Author's Rebuttal
We thank the referee for the constructive comments on our position paper. We clarify below that the contribution is primarily conceptual and argumentative, aimed at highlighting overlooked moderators rather than delivering a fully re-derived economic model. We respond to each major comment and indicate where revisions will be made.
read point-by-point responses
-
Referee: [Abstract] Abstract: The revision of the Gries and Naudé (2022) model is stated as redefining effective organizational determinants, but no explicit revised equations or equilibrium conditions are presented to show how the five factors are incorporated as primary determinants or to verify internal consistency of the modification.
Authors: We agree that the abstract and main text describe the revision at a conceptual level without presenting new equations. As a position paper, our intent is to argue that existing frameworks overlook key moderators rather than to produce a standalone modeling paper with full equilibrium derivations. The revision redefines determinants by treating the five factors as attenuators on the effective productivity term in the original partial equilibrium setup. We will revise the manuscript to include an expanded subsection that explicitly maps each factor to components of the Gries and Naudé model (e.g., how learning curve affects the effective labor input) while preserving the position-paper scope. revision: partial
-
Referee: [Discussion of the five factors] Discussion of the five factors: The identification of the five moderating factors (human resource composition, baseline capability, learning curve, incentives for fair use, flexibility of objectives) is presented as key to attenuating productivity gains, but without derivation from the base model or any validation steps, it is unclear whether these are load-bearing primary determinants.
Authors: The five factors are not mathematically derived from the base model; they are synthesized from organizational behavior and technology-adoption literature as the most direct moderators that can nullify AI-driven gains in practice. Their status as primary is argued on substantive grounds (each can independently block the translation from technical capability to output) rather than through formal derivation or empirical validation within this paper. We will add a short justification table with key citations for each factor and a paragraph explaining why they are treated as load-bearing in the revised framework, but we maintain that full validation would require a separate empirical study outside the position-paper format. revision: partial
Circularity Check
No significant circularity; draws on external model without reduction to self-inputs
full rationale
The paper is a position piece that identifies five moderating factors and states it revises the partial equilibrium model of Gries and Naudé (2022). No author overlap exists with the cited work, no equations appear in the supplied text, and no derivation reduces a claimed prediction or uniqueness result to a fitted parameter or self-citation by construction. The central argument rests on the assertion that the listed factors are overlooked, which is presented as interpretive rather than a closed mathematical loop. This satisfies the default expectation of self-contained non-circular reasoning.
Axiom & Free-Parameter Ledger
axioms (1)
- domain assumption The partial equilibrium model of Gries and Naudé (2022) can be revised to capture organizational AI productivity dynamics.
Reference graph
Works this paper leans on
-
[1]
and Levy, Frank and Murnane, Richard J
doi: 10.1162/003355303322552801. Baily, M., Byrne, D., Kane, A., and Soto, P. Generative AI at the crossroads: Light bulb, dynamo, or microscope? arXiv preprint arXiv:2505.14588, 2025. Bankins, S., Ocampo, A. C., Marrone, M., Restubog, S. L. D., and Woo, S. E. A multilevel review of artificial intelligence in organizations: Implications for organiza- tion...
-
[2]
ISBN 9780226613338. Brynjolfsson, E., Rock, D., and Syverson, C. The productiv- ity J-curve: How intangibles complement general purpose technologies.American Economic Journal: Macroeco- nomics, 13(1):333–372, 2021. Brynjolfsson, E., Li, D., and Raymond, L. R. Generative AI at work.The Quarterly Journal of Economics, 140(2): 889–942, 2025. doi: 10.1093/qje...
-
[3]
doi: 10.1186/s12651-022-00319-2. Gritsenko, D. and Wood, M. Algorithmic governance: A modes of governance approach.Regulation & Gover- nance, 16(1):45–62, 2022. J¨ohnk, J., Weißert, M., and Wyrtki, K. Ready or not, AI comes—an interview study of organizational AI readiness factors.Business & Information Systems Engineering, 63 (1):5–20, 2021. doi: 10.1007...
work page internal anchor Pith review Pith/arXiv arXiv doi:10.1186/s12651-022-00319-2 2022
-
[4]
Available at https://hai.stanford.edu/ ai-index/2024-ai-index-report. Mitra, B. and Craswell, N. An introduction to neural infor- mation retrieval.Foundations and Trends in Information Retrieval, 13(1):1–126, 2018. Ouyang, L., Wu, J., Jiang, X., Almeida, D., Wainwright, C., Mishkin, P., Zhang, C., Agarwal, S., Slama, K., Ray, A., et al. Training language ...
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.