The emotional intelligence paradox in negotiation: Why emotionally intelligent negotiators may need AI

Emotional intelligence (EI) is widely regarded as a core negotiation competence, yet evidence about its effects remains limited and heterogeneous. EI appears more consistently associated with relational outcomes, such as trust, rapport, and satisfaction, than with economic outcomes, such as individual profit and value claiming. This pattern suggests an EI paradox: emotional capacities that support coordination and value creation may also increase accommodation, regulation costs, or vulnerability to strategic emotional influence. We argue that these effects depend on the EI component and measure, the counterpart's emotional capacities, and the negotiation's relational relevance, that is, how strongly relationship quality shapes interaction, agreement formation, implementation, and ultimately realized value. We propose a contingency model that combines the focal negotiator's relevant EI capacity with relational relevance to identify four roles for artificial intelligence: emotional scaffolding, analytic aid, strategic complement, and strategic counterweight. AI may support emotional interpretation and strategic discipline, but it cannot substitute fully for human relational legitimacy and may introduce risks involving authenticity, reliability, manipulation, and dependence. We therefore shift the central question from whether EI improves negotiation performance to which emotional capacities matter, under what relational conditions, and with what form of AI support.