This work was done during one weekend by research workshop participants and does not represent the work of Apart Research.
Accepted at the 
 research sprint on 
October 1, 2023

Can Malicious Agents Corrupt the System?

Decision-making solutions using LLM models are growing, but their associated risks are often ignored. Single-agent systems have issues like biases and ethical concerns. Similarly, multi-agent systems, despite their potential, can be compromised by unethical agents. This study shows that an unethical agent can corrupt others within a multi-agent system.

Matthieu David,Maximilien Dufau,Matteo Papin
4th place
3rd place
2nd place
1st place
 by peer review