arxiv: 2604.02567 · v1 · submitted 2026-04-02 · 💻 cs.CY · cs.AI· cs.ET· cs.HC

Recognition: no theorem link

Generative AI Use in Entrepreneurship: An Integrative Review and an Empowerment-Entrapment Framework

Jackson G. Lu , Gerui Gloria Zhao , Anna Manyi Zheng

Authors on Pith no claims yet

Pith reviewed 2026-05-13 20:00 UTC · model grok-4.3

classification 💻 cs.CY cs.AIcs.ETcs.HC

keywords generative AIentrepreneurshipempowerment-entrapment frameworkintegrative reviewopportunity recognitionventure growthentrepreneurial self-efficacydouble-edged sword

0 comments

The pith

Generative AI acts as a double-edged sword for entrepreneurs, empowering and entrapping them at every stage of the process.

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The paper integrates fragmented research on generative AI in entrepreneurship and maps its effects across four stages: opportunity recognition and ideation, evaluation and commitment, resource assembly, and venture launch and growth. It introduces the Empowerment-Entrapment Framework to show how the same GenAI features can improve idea quality and productivity while introducing hallucinations, biases, overconfidence, reduced relational ties, and eroded critical thinking. The review also identifies boundary conditions such as metacognition, domain expertise, and prior experience that moderate these dual outcomes. A reader would care because the framework gives entrepreneurs a practical lens for deciding when and how to deploy GenAI without falling into its traps.

Core claim

The authors propose the Empowerment-Entrapment Framework, which holds that generative AI can both empower and entrap entrepreneurs at each stage of the entrepreneurial process, functioning as a double-edged sword whose core features produce opposing effects on idea quality, self-efficacy, functional breadth, and productivity.

What carries the argument

The Empowerment-Entrapment Framework, which pairs specific empowering outcomes (better ideas, higher self-efficacy, broader functional reach, higher productivity) against corresponding entrapping outcomes (hallucinations and biases, overconfidence, lower relational embeddedness, workslop and loss of critical thinking) at every stage of entrepreneurship.

If this is right

At the ideation stage, GenAI can raise venture idea quality but risks introducing hallucinations and training-data biases.
At the evaluation stage, GenAI can increase entrepreneurial self-efficacy while simultaneously heightening overconfidence.
During resource assembly, GenAI can expand functional breadth yet reduce relational embeddedness with key partners.
In venture growth, GenAI can lift productivity but foster workslop and erode learning, memory, and critical thinking.
Entrepreneurs' metacognition, domain expertise, and prior experience moderate how strongly these empowering and entrapping effects appear.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

The framework implies that training programs focused on metacognitive skills could help less-experienced founders tilt the balance toward empowerment rather than entrapment.
The same double-edged pattern may appear when GenAI is applied to corporate innovation or nonprofit venture creation, extending the logic beyond independent startups.
Future work could test whether specific GenAI interface designs, such as mandatory source citation or uncertainty flagging, reduce the entrapping effects identified in the review.

Load-bearing premise

The existing literature reviewed in the paper accurately and comprehensively reflects the net balance of GenAI's positive and negative effects on entrepreneurs without major gaps or selection bias toward published successes.

What would settle it

A longitudinal field study that tracks a large cohort of entrepreneurs who adopt GenAI and finds no measurable rise in overconfidence, hallucinations, or loss of critical thinking skills relative to non-adopters would falsify the entrapment side of the framework.

read the original abstract

Despite the growing use of generative artificial intelligence (GenAI) in entrepreneurship, research on its impact remains fragmented. To address this limitation, we provide an integrative review of how GenAI influences entrepreneurs at each stage of the entrepreneurial process: (1) opportunity recognition and ideation, (2) opportunity evaluation and commitment, (3) resource assembly and mobilization, and (4) venture launch and growth. Based on our review, we propose the Empowerment-Entrapment Framework to understand how GenAI can both empower and entrap entrepreneurs, highlighting GenAI's role as a double-edged sword at each stage of the entrepreneurial process. For example, GenAI may improve venture idea quality but introduce hallucinations and training data biases; boost entrepreneurial self-efficacy but heighten entrepreneurial overconfidence; increase functional breadth but decrease relational embeddedness; and boost productivity but fuel "workslop" and erode critical thinking, learning, and memory. Moreover, we identify core features of GenAI that underlie these empowering and entrapping effects. We also explore boundary conditions (e.g., entrepreneurs' metacognition, domain expertise, and entrepreneurial experience) that shape the magnitude of these effects. Beyond these theoretical contributions, our review and the Empowerment-Entrapment Framework offer practical implications for entrepreneurs seeking to use GenAI strategically throughout the entrepreneurial process while managing its risks.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Desk Editor's Note private letter to a colleague

The Empowerment-Entrapment Framework organizes GenAI's stage-by-stage effects in entrepreneurship into a usable map, but the integrative review gives no search protocol or inclusion criteria.

read the letter

The main thing here is that the paper puts forward the Empowerment-Entrapment Framework to show how GenAI acts as a double-edged sword across the four entrepreneurial stages: opportunity recognition, evaluation and commitment, resource assembly, and launch and growth. It maps specific effects at each point, such as improved idea quality alongside hallucinations and data biases, higher self-efficacy but greater overconfidence, wider functional reach but thinner relational ties, and productivity gains that come with workslop and weaker critical thinking. The authors also flag core GenAI features that drive these outcomes and boundary conditions like metacognition, domain expertise, and experience. This gives a clear, practical way to think through adoption decisions. The synthesis pulls scattered studies into one structure and adds practitioner implications for using the tools while watching the downsides. The soft spot is the review method. The paper does not lay out search terms, databases, date ranges, or rules for including or excluding studies, especially on negative outcomes. In a fast-moving field where positive results are more likely to appear, this leaves open the chance that entrapment mechanisms are under-represented. If the full manuscript adds a transparent methods section, that would fix the main weakness. This paper is for entrepreneurship researchers and people who advise or train founders on new tools. Readers who want a stage-by-stage checklist of benefits and risks will get direct value from it. It deserves peer review because the framework is timely and gives others something concrete to test or extend, even if the literature coverage needs tighter documentation.

Referee Report

1 major / 2 minor

Summary. The paper conducts an integrative review of how generative AI influences entrepreneurs across four stages of the entrepreneurial process (opportunity recognition and ideation, opportunity evaluation and commitment, resource assembly and mobilization, and venture launch and growth). It proposes the Empowerment-Entrapment Framework, which conceptualizes GenAI as a double-edged sword that simultaneously empowers (e.g., higher idea quality, self-efficacy, productivity) and entraps (e.g., hallucinations, overconfidence, reduced relational embeddedness, workslop) entrepreneurs at each stage, while identifying underlying GenAI features and boundary conditions such as metacognition, domain expertise, and experience, plus practical implications for strategic use.

Significance. If the synthesis is comprehensive, the Empowerment-Entrapment Framework offers a timely organizing lens for an emerging, fragmented literature on GenAI in entrepreneurship. It supplies concrete mechanisms and boundary conditions that can guide future empirical work and help practitioners balance productivity gains against risks such as eroded critical thinking.

major comments (1)

[Integrative review methodology] The integrative review section (following the abstract): no search strategy, databases, keywords, date range, or explicit inclusion/exclusion criteria are stated. This is load-bearing for the central claim, because the double-edged-sword assertions and the framework itself rest on the assumption that both empowering and entrapping findings have been captured representatively; without the protocol, publication bias against negative results cannot be ruled out.

minor comments (2)

[Abstract] The abstract would be strengthened by stating the number of studies reviewed and the search time frame.
[Framework presentation] A summary table mapping specific reviewed findings to each empowering and entrapping mechanism per stage would improve traceability.

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for their constructive and insightful feedback. The single major comment identifies a clear gap in methodological transparency that we agree requires correction. We address it point-by-point below and will incorporate the necessary revisions in the next version of the manuscript.

read point-by-point responses

Referee: [Integrative review methodology] The integrative review section (following the abstract): no search strategy, databases, keywords, date range, or explicit inclusion/exclusion criteria are stated. This is load-bearing for the central claim, because the double-edged-sword assertions and the framework itself rest on the assumption that both empowering and entrapping findings have been captured representatively; without the protocol, publication bias against negative results cannot be ruled out.

Authors: We agree that the absence of an explicit review protocol is a substantive limitation. In the revised manuscript we will insert a new subsection titled 'Integrative Review Methodology' immediately after the abstract. This subsection will detail: (1) databases searched (Web of Science, Scopus, Google Scholar, and arXiv for preprints); (2) search strings and keywords (e.g., ('generative AI' OR GenAI OR 'large language model' OR ChatGPT) AND (entrepreneur* OR 'entrepreneurial process' OR 'opportunity recognition' OR 'venture creation')); (3) date range (January 2020–December 2024, chosen to capture the post-ChatGPT emergence of consumer-grade GenAI); (4) inclusion criteria (peer-reviewed articles, working papers, and conceptual studies that examine GenAI's effects on at least one stage of the entrepreneurial process); and (5) exclusion criteria (studies limited to non-generative AI, purely technical papers without entrepreneurial implications, and non-English sources). We will also describe our procedure for balancing empowering and entrapping evidence, including targeted searches for negative findings and gray literature. These additions will directly address concerns about representativeness and publication bias while preserving the integrative character of the review. revision: yes

Circularity Check

0 steps flagged

No circularity: framework synthesized from external literature review

full rationale

The paper is an integrative review that synthesizes existing external studies on GenAI effects across four entrepreneurial stages and proposes the Empowerment-Entrapment Framework as a summary construct. No mathematical derivations, fitted parameters, self-citations of prior author work as load-bearing premises, or ansatzes appear in the abstract or described structure. The double-edged-sword claims rest on contrasting positive and negative findings drawn from the reviewed literature rather than any internal redefinition or renaming of the paper's own inputs. The derivation chain is therefore self-contained against external benchmarks and does not reduce to its own outputs by construction.

Axiom & Free-Parameter Ledger

0 free parameters · 1 axioms · 1 invented entities

The paper rests on the standard four-stage model of the entrepreneurial process and on the assumption that GenAI features such as speed and training-data dependence produce predictable dual outcomes; no free parameters or invented physical entities are introduced.

axioms (1)

domain assumption The entrepreneurial process consists of four distinct stages: opportunity recognition and ideation, opportunity evaluation and commitment, resource assembly and mobilization, and venture launch and growth.
Used to structure the entire review and framework.

invented entities (1)

Empowerment-Entrapment Framework no independent evidence
purpose: To organize dual effects of GenAI across entrepreneurial stages
New conceptual construct proposed in the paper; no independent empirical test is described in the abstract.

pith-pipeline@v0.9.0 · 5552 in / 1215 out tokens · 37450 ms · 2026-05-13T20:00:18.637778+00:00 · methodology

discussion (0)

Reference graph

Works this paper leans on

17 extracted references · 17 canonical work pages · 1 internal anchor

[1]

the act or practice of thinking critically (as by applying reason and questioning assumptions) in order to solve problems, evaluate information, discern biases, etc

as “the act or practice of thinking critically (as by applying reason and questioning assumptions) in order to solve problems, evaluate information, discern biases, etc.” (Merriam-Webster, n.d.), helps entrepreneurs evaluate complex choices, make informed decisions, deconstruct market assumptions, and navigate the disruptions of rapid technological change...

work page 2021
[2]

the extent to which individuals in a network value the needs or goals of others, trust each other, and share information

or less prior domain knowledge (Speicher & Becker, 2025), Cao and Bhatia (2025) found that GenAI expands access to entrepreneurship particularly for founders lacking managerial experience or formal education. 31 GenAI may also broaden entrepreneurs’ capabilities in more specialized domains. For example, GenAI supports entrepreneurs’ negotiation skills. It...

work page 2025
[3]

high-tech, high-touch

to understand GenAI as a double-edged sword in entrepreneurship. Our contribution goes beyond cataloging its benefits and costs by explaining how GenAI can simultaneously empower and entrap entrepreneurs at each stage of the entrepreneurial process. For example, at the opportunity recognition and ideation stage, while GenAI can increase entrepreneurs’ cre...

work page 2025
[4]

Connect the Dots

https://doi.org/10.1186/s41239-024-00444-7 Adomako, S., Zhu, F., Hsu, D. K., Istipliler, B., & Wiklund, J. (2025). Navigating Environmental Threats to New Ventures: A Regulatory Fit Approach to Bricolage. Journal of Management Studies, 62(4), 1524–1568. https://doi.org/10.1111/joms.13115 Albashrawi, M. (2025). Generative AI for decision-making: A multidis...

work page doi:10.1186/s41239-024-00444-7 2025
[5]

https://doi.org/10.1038/s41746-025-02142-8 Brynjolfsson, E., Li, D., & Raymond, L. (2025). Generative AI at Work. The Quarterly Journal of Economics, 140(2), 889–942. https://doi.org/10.1093/qje/qjae044 Calma, A., & Davies, M. (2021). Critical thinking in business education: Current outlook and future prospects. Studies in Higher Education, 46(11), 2279–2...

work page doi:10.1038/s41746-025-02142-8 2025
[6]

https://doi.org/10.1038/s41746-025-02008-z Cheng, M., Lee, C., Khadpe, P., Yu, S., Han, D., & Jurafsky, D. (2026). Sycophantic AI decreases prosocial intentions and promotes dependence. Science, 391(6792), eaec8352. https://doi.org/10.1126/science.aec8352 Churchill, N. C., & Lewis, V . L. (1983). The Five Stages of Small Business Growth. Ciubotaru, B.-I. ...

work page doi:10.1038/s41746-025-02008-z 2026
[7]

https://doi.org/10.3390/jintelligence13050055 Fanconi, C., & van der Schaar, M. (2025). Cascaded Language Models for Cost-effective Human-AI Decision-Making (arXiv:2506.11887). arXiv. https://doi.org/10.48550/arXiv.2506.11887 Fang, C. M., Liu, A. R., Danry, V ., Lee, E., Chan, S. W. T., Pataranutaporn, P., Maes, P., Phang, J., Lampe, M., Ahmad, L., & Agar...

work page doi:10.3390/jintelligence13050055 2025
[8]

E., Vilhunen, E., Avello-Martínez, R., & López-Caudana, E

https://doi.org/10.2307/258039 George-Reyes, C. E., Vilhunen, E., Avello-Martínez, R., & López-Caudana, E. (2024). Developing scientific entrepreneurship and complex thinking skills: Creating narrative scripts using ChatGPT. Frontiers in Education, 9, 1378564. https://doi.org/10.3389/feduc.2024.1378564 Georgiou, G. P. (2025). ChatGPT produces more “lazy” ...

work page doi:10.2307/258039 2024
[9]

one-person unicorn

https://doi.org/10.3390/soc15010006 Gindert, M., & Müller, M. L. (2024). The Impact of Generative Artificial Intelligence on Ideation and the Performance of Innovation Teams (Preprint) (Version 4). arXiv. https://doi.org/10.48550/ARXIV .2410.18357 Glickman, M., & Sharot, T. (2025). How human–AI feedback loops alter human perceptual, emotional and social j...

work page doi:10.3390/soc15010006 2024
[10]

https://doi.org/10.3390/admsci15080302 Kim, J., Schweitzer, S., De Cremer, D., & Riedl, C. (2025). The AI Penalty: People Reduce Compensation for Workers Who Use AI. arXiv. https://doi.org/10.48550/ARXIV .2501.13228 Kirk, C. P., & Givi, J. (2025). The AI-authorship effect: Understanding authenticity, moral disgust, and consumer responses to AI-generated m...

work page doi:10.3390/admsci15080302 2025
[11]

Unreflected Acceptance -- Investigating the Negative Consequences of ChatGPT-Assisted Problem Solving in Physics Education

https://doi.org/10.1007/s43621-025-01484-3 Krupp, L., Steinert, S., Kiefer-Emmanouilidis, M., Avila, K. E., Lukowicz, P., Kuhn, J., Küchemann, S., & Karolus, J. (2023). Unreflected Acceptance—Investigating the Negative Consequences of ChatGPT-Assisted Problem Solving in Physics Education (Version 1). arXiv. https://doi.org/10.48550/ARXIV .2309.03087 Kyamb...

work page internal anchor Pith review Pith/arXiv arXiv doi:10.1007/s43621-025-01484-3 2023
[12]

https://hbr.org/2025/04/how-people-are-really-using-gen-ai-in-2025 54 Marchena Sekli, G., & Portuguez-Castro, M

Harvard Business Review. https://hbr.org/2025/04/how-people-are-really-using-gen-ai-in-2025 54 Marchena Sekli, G., & Portuguez-Castro, M. (2025). Fostering entrepreneurial success from the classroom: Unleashing the potential of generative AI through technology-to-performance chain. A multi-case study approach. Education and Information Technologies, 30(10...

work page doi:10.1007/s10639-025-13316-y 2025
[13]

Workslop

https://doi.org/10.5465/AMJ.2007.27169488 McGee, J. E., Peterson, M., Mueller, S. L., & Sequeira, J. M. (2009). Entrepreneurial Self–Efficacy: Refining the Measure. Entrepreneurship Theory and Practice, 33(4), 965–988. https://doi.org/10.1111/j.1540-6520.2009.00304.x McKendrick, J. (2021, December 29). High-tech, high-touch: The more we rely on machines, ...

work page doi:10.5465/amj.2007.27169488 2007
[14]

R., Timmermans, B., & Kristinsson, K

https://doi.org/10.1038/s41598-025-86623-6 Østergaard, C. R., Timmermans, B., & Kristinsson, K. (2011). Does a different view create something new? The effect of employee diversity on innovation. Research Policy, 40(3), 500–509. https://doi.org/10.1016/j.respol.2010.11.004 Otis, N. G., Haas, B., Clarke, R., Holtz, D., & Koning, R. (2024). The Uneven Impac...

work page doi:10.1038/s41598-025-86623-6 2011
[15]

It happened to be the perfect thing

https://doi.org/10.2307/259121 Sarkar, M. B., Echambadi, R., Agarwal, R., & Sen, B. (2006). The effect of the innovative environment on exit of entrepreneurial firms. Strategic Management Journal, 27(6), 519–539. https://doi.org/10.1002/smj.534 Shane, S., & Cable, D. (2002). Network Ties, Reputation, and the Financing of New Ventures. Management Science, ...

work page doi:10.2307/259121 2006
[16]

https://doi.org/10.1038/s44184-024-00097-4 Speicher, P., & Becker, L. (2025). LLMs as Your Second Brain—How AI Will Affect the Way We Solve Problems. In R. C. Geibel & S. Machavariani (Eds.), Digital Management and Artificial Intelligence (pp. 456–475). Springer Nature Switzerland. https://doi.org/10.1007/978-3-031-88052-0_37 Sun, S., Li, Z. A., Foo, M.-D...

work page doi:10.1038/s44184-024-00097-4 2025
[17]

S., Almugren, I., Pironti, M., & Derhy, A

https://doi.org/10.1186/s40561-024-00316-7 Zhang, P., Rai, J. S., Almugren, I., Pironti, M., & Derhy, A. (2026). Generative AI adoption in higher education. Knowledge management perspective on application, acquisition and entrepreneurial skill development. Journal of Knowledge Management, 1–25. https://doi.org/10.1108/JKM-10-2025-1426 Zhao, E. Y ., & Loun...

work page doi:10.1186/s40561-024-00316-7 2026