Back
A Game-Theoretic Model of Global AI Development Race
webpreprints.org·preprints.org/manuscript/202409.1287/v1
A formal preprint using game theory to model AI race dynamics; relevant for researchers and policymakers studying how competitive pressures between AI developers may undermine safety and what governance interventions could help.
Metadata
Importance: 52/100working paperanalysis
Summary
This preprint applies game-theoretic frameworks to model the competitive dynamics of global AI development, analyzing how nations and actors make strategic decisions under competitive pressure. It examines how race dynamics may undermine safety incentives and explores conditions under which coordination or defection is rational. The work aims to inform governance strategies for managing AI competition.
Key Points
- •Models the AI development race as a strategic game between multiple actors (nations/firms), capturing incentives to prioritize speed over safety
- •Analyzes equilibrium outcomes where competitive pressure leads to collectively suboptimal results, such as reduced safety investment
- •Explores conditions under which international coordination mechanisms could shift actors toward safer development trajectories
- •Connects formal game-theoretic results to real-world AI governance challenges including treaties, standards, and compute controls
- •Highlights the role of information asymmetry and uncertainty in exacerbating race-to-the-bottom dynamics
Cited by 1 page
| Page | Type | Quality |
|---|---|---|
| Multipolar Trap (AI Development) | Risk | 91.0 |
Cached Content Preview
HTTP 200Fetched Mar 20, 20264 KB
Cite
Add to My List
Share Comments
Download PDF
Version 1
Submitted:
16 September 2024
Posted:
18 September 2024
You are already at the latest version
Subscription
Notify me about updates to this article or when a peer-reviewed version is published.
Subscribe
Preprint
Article
This version is not peer-reviewed.
[Richard Murdoch Montgomery](https://sciprofiles.com/profile/author/cXdUUjZYbFZBSEduS3dQbnZSdmhndG5lTjBmbXlLT0hlT2pGZTRaY1lpQT0=) \*
[Richard Murdoch Montgomery](https://sciprofiles.com/profile/author/cXdUUjZYbFZBSEduS3dQbnZSdmhndG5lTjBmbXlLT0hlT2pGZTRaY1lpQT0=) \*
Show more
Version 1
Submitted:
16 September 2024
Posted:
18 September 2024
You are already at the latest version
###### Abstract
###### Keywords:
;
;
;
;
;
;
;
;
;
###### Subject:
[Computer Science and Mathematics](https://www.preprints.org/subject/browse/computer-science-and-mathematics) \- [Artificial Intelligence and Machine Learning](https://www.preprints.org/subject/browse/computer-science-and-mathematics/artificial-intelligence-and-machine-learning)
Copyright: This open access article is published under a [Creative Commons CC BY 4.0 license](https://creativecommons.org/licenses/by/4.0/), which permit the free download, distribution, and reuse, provided that the author and preprint are cited in any reuse.
[](https://www.preprints.org/)
Preprints.org is a free preprint server supported by MDPI in Basel, Switzerland.
[](https://www.facebook.com/Preprints.org/)[](https://twitter.com/Preprints_org)[](https://www.linkedin.com/showcase/preprints)
[](https://bsky.app/profile/preprints.bsky.social)
[Contact Us](mailto:info@preprints.org) RSS
##### MDPI Initiatives
- [SciProfiles](https://sciprofiles.com/)
- [Sciforum](https://sciforum.net/)
- [Encyclopedia](https://encyclopedia.pub/)
- [MDPI Books](https://www.mdpi.com/books)
- [Scilit](https://www.scilit.com/)
- [Proceedings](https://www.mdpi.com/about/proceedings)
- [JAMS](https://jams.pub/)
##### Important Links
- [Advisory Board](https://www.preprints.org/advisory-board)
- [Collections](https://www.preprints.org/collection)
- [How It Works](https://www.preprints.org/about?scrollTo=works)
- [Preprints Friendly Journals](https://www.preprints.org/friendly-journals)
- [Reading List](https://www.preprints.org/reading-list)
- [News](https://www.preprints.org/news)
- [Statistics](https://www.preprints.org/statistics)
- [2023 Popular Award](https://www.preprints.org/activity/award/announcement)
- [2025 Book
... (truncated, 4 KB total)Resource ID:
99d4add39a446e74 | Stable ID: YzI0OWY0Mj