Imagine two AI agents haggling over the price of a used car. One’s a hardball negotiator, the other’s a pushover. But what if these weren’t just quirks of their programming, but reflections of distinct ‘personalities’? That’s the fascinating premise explored by researchers at Kyoto University, who built a negotiation simulator using large language models (LLMs) imbued with different personality traits. They used the Big Five personality framework—openness, conscientiousness, extraversion, agreeableness, and neuroticism—to craft unique profiles for each AI agent. These digital personalities weren’t just for show. The researchers found that an AI’s personality significantly influenced how it negotiated, mirroring patterns seen in human interactions. Agreeable AI, for example, tended to concede more readily, striking deals quickly but often at less advantageous terms. Conscientious AI, on the other hand, were sticklers, holding their ground and often securing better deals. Interestingly, extroverted AI negotiators proved more successful at closing deals, a finding that echoes observations in human sales. The research also revealed how different strategies play out in AI negotiations. Cooperative tactics like accommodation and concession led to fairer outcomes for both parties. However, aggressive strategies, often employed by disagreeable AI, resulted in less balanced deals. This work offers a compelling glimpse into the complex interplay of personality, strategy, and outcome in a simulated negotiation. It raises intriguing questions about the future of AI and its role in everything from automated customer service to high-stakes business deals. As AI systems become increasingly integrated into our lives, understanding the subtleties of their ‘behavior’ becomes crucial. Could these insights help us build more effective and ethical AI negotiators? Or perhaps, they hold a mirror to ourselves, revealing hidden biases in our own negotiation styles.