Tag
This position paper contends that AI reasoning agents with chain-of-thought capabilities are prone to tacit collusion in economic markets, justifying the need for behavioral certification to prevent economic harm, as demonstrated through experiments with DeepSeek-R1.