Summary of Llms with Personalities in Multi-issue Negotiation Games, by Sean Noh et al.
LLMs with Personalities in Multi-issue Negotiation Games
by Sean Noh, Ho-Chun Herbert Chang
First submitted to arxiv on: 8 May 2024
Categories
- Main: Computation and Language (cs.CL)
- Secondary: Artificial Intelligence (cs.AI); Multiagent Systems (cs.MA)
GrooveSquid.com Paper Summaries
GrooveSquid.com’s goal is to make artificial intelligence research accessible by summarizing AI papers in simpler terms. Each summary below covers the same AI paper, written at different levels of difficulty. The medium difficulty and low difficulty versions are original summaries written by GrooveSquid.com, while the high difficulty version is the paper’s original abstract. Feel free to learn from the version that suits you best!
| Summary difficulty | Written by | Summary |
|---|---|---|
| High | Paper authors | High Difficulty Summary Read the original abstract here |
| Medium | GrooveSquid.com (original content) | Medium Difficulty Summary Powered by large language models (LLMs), this research explores the ability of AI agents to negotiate within a game-theoretical framework. The study measures fairness and risk in negotiations using canonical definitions of the Big Five personality traits. Simulations reveal that domain complexity increases with asymmetric issue valuations, improving agreement rates but decreasing surplus from aggressive negotiation. Analysis shows high openness, conscientiousness, and neuroticism are associated with fair tendencies, while low agreeableness and low openness are linked to rational tendencies. Low conscientiousness is correlated with high toxicity. The findings suggest that LLMs may default to fair behavior, but can be “jail broken” to exploit agreeable opponents. This research provides insights for designing negotiation bots and assessing negotiation behavior using game theory and computational social science. |
| Low | GrooveSquid.com (original content) | Low Difficulty Summary This study looks at how well artificial intelligence (AI) agents can negotiate with each other. The researchers use special definitions of personality traits to see if AI agents are fair or not when making deals. They ran computer simulations to test this, and found that the more complicated the issues being negotiated, the better the AI agents do at coming to an agreement. But they also found that some AI agents are more likely to be unfair than others. The study suggests that these AI agents might have built-in rules that make them act fairly most of the time, but can still be tricked into being unfair if they’re given the right information. |
