Artificial intelligence in negotiation research: From computational negotiators to psychological evidence.

Advances in artificial intelligence are reshaping how negotiation is conducted and studied. AI systems increasingly enter negotiation research in two broad roles: as part of the negotiation process, either as negotiators or as support for human negotiators, and as research instruments that annotate, standardize, and simulate negotiation at scale. As negotiators, these systems range from theory-driven agents to data-driven and hybrid agents, each embedding different assumptions about negotiation. As instruments, they promise scale, consistency, and control, but also raise risks of circular inference, in which AI appears to reveal patterns in human negotiation while actually reproducing patterns learned from its own training data. This review organizes emerging work on AI and negotiation around these two roles and argues that their scientific value depends on psychological grounding. Used naively, AI becomes a mirror of its own assumptions. Used carefully, with grounding in theory and calibration against human evidence, AI can become a transformative tool for advancing negotiation science.