Competitive Market Behavior of LLMs
When AI agents replace humans in markets, efficiency breaks down
When researchers replaced human traders with AI language models in a classic economic experiment, the markets failed to reach equilibrium and produced worse resource allocation than human-run markets. The AI agents showed wildly different trading behaviors depending on which model they used, and analysis of their reasoning revealed they switched from strategic thinking to impulsive "let's just trade now" urgency.
As companies deploy large language models as autonomous economic agents—in trading, bidding systems, and marketplace negotiations—these results show that markets designed for human behavior may malfunction with AI. The inefficient allocations mean potential losses for buyers and sellers, and the unpredictability across different AI models creates risk for anyone building systems that mix human and AI traders.