MaMa: A Game-Theoretic Approach for Designing Safe Agentic Systems
Jonathan Nöther, Adish Singla, Goran Radanovic · 2026
Read MaMa: A Game-Theoretic Approach for Designing Safe Agentic Systems by Jonathan Nöther, Adish Singla, Goran Radanovic (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2602-04431.txt.
Description
Design the agent team assuming some members will be turned against it. Jonathan Nöther and 2 co-authors cast that paranoia as a game in a 2026 paper. Their MaMa method pits a system designer against a meta-adversary that compromises a subset of agents to cut safety, echoing Stackelberg security games.
The designer proposes architectures in rounds; language-model-driven adversarial search returns the strongest attacks found as feedback. Across diverse environments, MaMa-designed systems defended against worst-case attacks while keeping performance on par with systems built without adversaries. Built-in red-teaming inside the design loop beats bolting safety on later.
Catalog record
| CREATOR | Jonathan Nöther, Adish Singla, Goran Radanovic |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 78 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-04431