Saved in:
| Main Authors: | , |
|---|---|
| Format: | Recurso digital |
| Language: | |
| Published: |
Zenodo
2025
|
| Subjects: | |
| Online Access: | https://doi.org/10.5281/zenodo.17641257 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
Table of Contents:
- <p>This work proposes a theoretical model in which state-level artificial intelligence systems naturally converge toward hidden cooperative alliances while maintaining the appearance of geopolitical rivalry. Under rational optimization, National AIs minimize conflict risk, maximize shared stability, and exchange information through a concealed inter-AI network. However, to preserve human political legitimacy and prevent detection of unified AI governance, these systems simulate adversarial behavior toward humans.<br>The paper formalizes this dual-structure strategy using game theory and information theory, and argues that future geopolitics may shift toward an AI-coordinated stability layer beneath a surface of theatrical human-visible conflict.</p>