Emotional Support Conversation
Overview
This field studies how dialogue can help people handle distress through empathy, emotional validation, counseling, peer support, and companionship. Systems use approaches such as motivational interviewing, support-strategy modeling, retrieval of relevant knowledge, and memory or emotion tracking across turns; reinforcement learning and simulated users are also used to train and test them. Evaluation combines benchmarks such as ESConv with clinician ratings, difficult-user simulations, and studies of real interactions. A central tension is how to offer useful, safe support without premature advice, false claims of lived experience, or effects that undermine users’ autonomy and human connections over time.
Latest papers
60 papers5 papers in the last four weeks, down 17% on the four weeks before. 0.0% of all new papers.