CoChat
Sign Up
PreprintCited by 3Digital Platforms and EconomicsTechnology's Impact on MediaEthics and Social Impacts of AI

The DOI matches its DataCite record; 1 of the 1 references it cites were found in a registry or an index.

Human vs. Generative AI in Content Creation Competition: Symbiosis or Conflict?

Open access
Yao, Fan, Li, Chuanhao, Nekipelov, Denis, Wang, Hongning, Xu, Haifeng·2024·arXiv
DOI 10.48550/arxiv.2402.15467arXiv 2402.15467
Open this paper in CoChat — free accountOpen PDF

Abstract

The advent of generative AI (GenAI) technology produces transformative impact on the content creation landscape, offering alternative approaches to produce diverse, high-quality content across media, thereby reshaping online ecosystems but also raising concerns about market over-saturation and the potential marginalization of human creativity. Our work introduces a competition model generalized from the Tullock contest to analyze the tension between human creators and GenAI. Our theory and simulations suggest that despite challenges, a stable equilibrium between human and AI-generated content is possible. Our work contributes to understanding the competitive dynamics in the content creation industry, offering insights into the future interplay between human creativity and technological advancements in GenAI.

Overview

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

DataCite · 10.48550/arxiv.2402.15467

Cites

1 of 1 references verified0 citation errors · 0 not checkable

Statistics

Verification

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Citations

Cites

  • Scaling Laws for Neural Language Models

    Kaplan, J., McCandlish, S., Henighan, T., Brown, T. B., Chess, B., Child, R., Gray, S., Radford, A., Wu, J., & Amodei, D. (2020) · arXiv · doi:10.48550/arxiv.2001.08361 · Cited by 9,550 · cited 1×

    “Our assumption here echos the “scaling law” in recent studies of large language models (Kaplan et al., 2020), where it is observed that the test loss scales as a power-law with model size and data size.”
    Method

Research works printed but not linked: 50

  1. Krishna Acharya, Varun Vangala, Jingyan Wang, and Juba Ziani. Producers equilibria and dynamics in engagement-driven recommender systems. arXiv preprint arXiv:2401.16641, 2024. · arXiv
  2. Arpit Agarwal and William Brown. Diversified recommendations for agents with adaptive preferences. arXiv preprint arXiv:2210.07773, 2022. · arXiv
  3. Nick Arnosti and S Matthew Weinberg. Bitcoin: A natural oligopoly. Management Science, 68(7):4755-4771, 2022.
  4. Omer Ben-Porat and Moshe Tennenholtz. Shapley facility location games. In International Conference on Web and Internet Economics, pages 58-73. Springer, 2017.
  5. Omer Ben-Porat and Moshe Tennenholtz. A game-theoretic approach to recommendation systems with strategic content providers. Advances in Neural Information Processing Systems, 31, 2018.
  6. Omer Ben-Porat, Itay Rosenberg, and Moshe Tennenholtz. Content provider dynamics and coordination in recommendation ecosystems. Advances in Neural Information Processing Systems, 33:18931-18941, 2020.
  7. Quentin Bertrand, Avishek Joey Bose, Alexandre Duplessis, Marco Jiralerspong, and Gau- thier Gidel. On the stability of iterative retraining of generative models on their own data. arXiv preprint arXiv:2310.00429, 2023. · arXiv
  8. Craig Boutilier, Martin Mladenov, and Guy Tennenholtz. Modeling recommender ecosys- tems: Research challenges at the intersection of mechanism design, reinforcement learning and generative models. arXiv preprint arXiv:2309.06375, 2023. · arXiv
  9. Mario Bravo, David Leslie, and Panayotis Mertikopoulos. Bandit learning in concave n- person games. Advances in Neural Information Processing Systems, 31, 2018.
  10. Martin Briesch, Dominik Sobania, and Franz Rothlauf. Large language models suffer from their own output: An analysis of the self-consuming training loop. arXiv preprint arXiv:2311.16822, 2023. · arXiv
  11. Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. Lan- guage models are few-shot learners. Advances in neural information processing systems, 33:1877-1901, 2020.
  12. Brian S Butler, Patrick J Bateman, Peter H Gray, and E Ilana Diamant. An attraction- selection-attrition theory of online community size and resilience. Mis Quarterly, 38(3): 699-729, 2014.
  13. Yang Cai and Weiqiang Zheng. Doubly optimal no-regret learning in monotone games. arXiv preprint arXiv:2301.13120, 2023. · arXiv
  14. Sarah Dean and Jamie Morgenstern. Preference dynamics under personalized recommen- dations. In Proceedings of the 23rd ACM Conference on Economics and Computation, pages 795-816, 2022.
  15. Sarah Dean, Evan Dong, Meena Jagadeesan, and Liu Leqi. Recommender systems as dynamical systems: Interactions with viewers and creators. In Workshop on Recommen- dation Ecosystems: Modeling, Optimization and Incentive Design, 2024.
  16. Gerard Debreu. A social equilibrium existence theorem. Proceedings of the National Academy of Sciences, 38(10):886-893, 1952.
  17. Anil R Doshi and Oliver Hauser. Generative ai enhances individual creativity but reduces the collective diversity of novel content. Available at SSRN, 2023.
  18. Ziv Epstein, Aaron Hertzmann, Investigators of Human Creativity, Memo Akten, Hany Farid, Jessica Fjeld, Morgan R Frank, Matthew Groh, Laura Herman, Neil Leach, et al. Art and the science of generative ai. Science, 380(6650):1110-1111, 2023.
  19. Eyal Even-Dar, Yishay Mansour, and Uri Nadav. On the convergence of regret minimization dynamics in concave games. In Proceedings of the forty-first annual ACM symposium on Theory of computing, pages 523-532, 2009.
  20. Christian Ewerhart. The lottery contest is a best-response potential game. Eco- nomics Letters, 155:168-171, 2017. ISSN 0165-1765. doi: https://doi.org/10.1016/ j.econlet.2017.03.030. URL https://www.sciencedirect.com/science/article/pii/ S0165176517301325. · DOI
  21. Ky Fan. Fixed-point and minimax theorems in locally convex topological linear spaces. Proceedings of the National Academy of Sciences, 38(2):121-126, 1952.
  22. Abheek Ghosh. Best-response dynamics in tullock contests with convex costs. arXiv preprint arXiv:2310.03528, 2023. · arXiv
  23. Irving L Glicksberg. A further generalization of the kakutani fixed theorem, with application to nash equilibrium points. Proceedings of the American Mathematical Society, 3(1):170- 174, 1952.
  24. Jonathan Ho, Ajay Jain, and Pieter Abbeel. Denoising diffusion probabilistic models. Ad- vances in neural information processing systems, 33:6840-6851, 2020.
  25. Jiri Hron, Karl Krauth, Michael Jordan, Niki Kilbertus, and Sarah Dean. Modeling con- tent creator incentives on algorithm-curated platforms. In The Eleventh International Conference on Learning Representations, 2022.
  26. Xinyan Hu, Meena Jagadeesan, Michael I Jordan, and Jacob Steinhard. Incentivizing high- quality content in online recommender systems. arXiv preprint arXiv:2306.07479, 2023. · arXiv
  27. Nicole Immorlica, Meena Jagadeesan, and Brendan Lucier. Clickbait vs. quality: How engagement-based optimization shapes the content landscape in online platforms. In Proceedings of the ACM Web Conference 2024, 2024.
  28. Meena Jagadeesan, Nikhil Garg, and Jacob Steinhardt. Supply-side equilibria in recom- mender systems. arXiv preprint arXiv:2206.13489, 2022. · arXiv
  29. Meena Jagadeesan, Nikhil Garg, and Jacob Steinhardt. Supply-side equilibria in recom- mender systems. Advances in Neural Information Processing Systems, 2023.
  30. Jacob D Leshno and Philipp Strack. Bitcoin: An axiomatic approach and an impossibility theorem. American Economic Review: Insights, 2(3):269-286, 2020.
  31. Martin Mladenov, Elliot Creager, Omer Ben-Porat, Kevin Swersky, Richard Zemel, and Craig Boutilier. Optimizing long-term social welfare in recommender systems: A con- strained matching approach. In International Conference on Machine Learning, pages 6987-6998. PMLR, 2020.
  32. Dennis C Mueller. Public choice III. Cambridge University Press, 2003.
  33. John F Nash Jr. Equilibrium points in n-person games. Proceedings of the national academy of sciences, 36(1):48-49, 1950.
  34. J David Pérez-Castrillo and Thierry Verdier. A general analysis of rent-seeking games. Public choice, 73(3):335-350, 1992.
  35. J Ben Rosen. Existence and uniqueness of equilibrium points for concave n-person games. Econometrica: Journal of the Econometric Society, pages 520-534, 1965.
  36. George Selgin. Gresham's law. Handbook of the History of Money and Currency, pages 199-219, 2020.
  37. Ronald William Shephard. Theory of cost and production functions. Princeton University Press, 2015.
  38. Stefan Szymanski. The economic design of sporting contests. Journal of economic literature, 41(4):1137-1187, 2003.
  39. Wondwesen Tafesse and Mumin Dayan. Content creators' participation in the creator economy: Examining the effect of creators' content sharing frequency on user engagement behavior on digital platforms. Journal of Retailing and Consumer Services, 73:103357, 2023.
  40. Tatiana Tatarenko and Maryam Kamgarpour. Bandit learning in convex non-strictly mono- tone games. arXiv preprint arXiv:2009.04258, 2020. · arXiv
  41. Gordon Tullock. Efficient rent seeking. In Toward a theory of the rent-seeking society. College Station, TX: Texas A&M University Press, 1980.
  42. Krzysztof Wach, Cong Doanh Duong, Joanna Ejdys, Rūta Kazlauskaitė, Pawel Korzyn- ski, Grzegorz Mazurek, Joanna Paliszkiewicz, and Ewa Ziemba. The dark side of gen- erative artificial intelligence: A critical analysis of controversies and risks of chatgpt. Entrepreneurial Business and Economics Review, 11(2):7-30, 2023.
  43. Risqo Wahid, Joel Mero, and Paavo Ritala. Written by chatgpt, illustrated by midjourney: generative ai for content marketing. Asia Pacific Journal of Marketing and Logistics, 35 (8):1813-1822, 2023.
  44. Fan Yao, Chuanhao Li, Denis Nekipelov, Hongning Wang, and Haifeng Xu. Learning from a learning user for optimal recommendations. In International Conference on Machine Learning, pages 25382-25406.
  45. PMLR, 2022a.
  46. Fan Yao, Chuanhao Li, Denis Nekipelov, Hongning Wang, and Haifeng Xu. Learning the optimal recommendation from explorative users. In Proceedings of the AAAI Conference on Artificial Intelligence, volume 36, pages 9457-9465, 2022b.
  47. Fan Yao, Chuanhao Li, Denis Nekipelov, Hongning Wang, and Haifeng Xu. How bad is top-k recommendation under competing content creators? In International Conference on Machine Learning, pages 39674-39701.
  48. PMLR, 2023a.
  49. Fan Yao, Chuanhao Li, Karthik Abinav Sankararaman, Yiming Liao, Yan Zhu, Qifan Wang, Hongning Wang, and Haifeng Xu. Rethinking incentives in recommender systems: Are monotone rewards always beneficial? Advances in Neural Information Processing Sys- tems, 2023b.
  50. Banghua Zhu, Sai Praneeth Karimireddy, Jiantao Jiao, and Michael I Jordan. Online learning in a creator economy. arXiv preprint arXiv:2305.11381, 2023. Denote (0.176, 0.118) = (x * 11 , x * 12 ) and (0.124, 0.085) = (x * 21 , x * 22 ). Since the con- structed G IN instance is symmetric, starting from ((x * 11 , x * 12 ), ⊥) we also have the following best-response chain ((x * 21 , x * 22 ), ⊥) → ((x * 21 , x * 22 ), (x * 11 , x * 12 )) → (⊥, (x * 11 , x * 12 )) → (⊥, (x * 21 , x * 22 )). · arXiv

Cited by

Sorted by citations

  • Foundations of GenIR

    Ai, Q., Zhan, J., & Liu, Y. (2025) · arXiv · doi:10.48550/arxiv.2501.02842

    “…and crush everything in the field [129], as it has almost happened in NLP. Interestingly, in simulated social experiments on human-AI competitions, Yao et al. [153] find that, if human producers don’t extend their capacities with the help of generative AI, they will eventually be “replaced” by AI.”

See all 3

Metadata

Authors
Yao, F., Li, C., Nekipelov, D., Wang, H., & Xu, H.
Affiliations
  • Yale University (US)
  • Tsinghua University (CN)
  • University of Chicago (US)
ORCID
Venue
arXiv
Publisher
arXiv
Published
2024
ISSN
2331-8422 (ISSN-L)
DOI
10.48550/arxiv.2402.15467
arXiv
2402.15467
OpenAlex
W4392182550
Cited by
3
References
1
Citations by year
2026: 1 · 2025: 1 · 2024: 1
Type
posted content
Language
en
Open access
Licence
arxiv.org/licenses/nonexclusive-distrib/1.0/
Locations
  • arXiv (Cornell University) (submittedVersion)
  • arXiv (Cornell University)
Funders
  • National Science Foundation (2213700, 2303372, 2128019)
  • Office of Naval Research (N00014-23-1-2802, N00014)
  • Army Research Office (W911NF23-1-0030, W911NF)
Topics
  • Social Sciences › Business, Management and Accounting › Strategy and Management › Digital Platforms and Economics
  • Social Sciences › Communication › Technology's Impact on Media
  • Physical Sciences › Computer Science › Artificial Intelligence › Ethics and Social Impacts of AI
Keywords
generative AI, content creation, Tullock contest, contest theory, equilibrium, market saturation, human creativity, AI-generated content

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review