Calendar of events

Title: A Digital Shoulder to Cry On: Understanding Why Large Language Models Can Surpass Humans as Extrinsic Interpersonal Emotion Regulators

Abstract: Does artificial intelligence (AI) provide emotional support as effective as humans? In this talk, I present research examining when and why large language models (LLMs) are preferred as extrinsic emotion regulators (EER) in non-clinical contexts. Across three studies, we systematically compared the regulatory effects and perceived effectiveness of LLM- versus human-generated comforting messages.

The first part of Study 1 (N = 279) used thematic analysis to show that LLM-generated responses largely mirrored human EER strategies, with some discrepancies between different LLMs. The second part of Study 1 (N = 309) demonstrated that LLMs generally outperformed human regulators with no differences observed for the sadness scenarios and the emotional response of fear (outcome). Study 2 (N = 196) showed again some LLM advantage in some scenarios and outcomes but emotional validation (i.e., acknowledgement of the target’s emotional response) did not explain the higher emotional improvement achieved by LLMs. Study 3 (N = 188) identified actionable support (i.e., specific and implementable regulatory tactics) as a key factor explaining LLM’s regulatory advantage in the previous studies. These findings suggest that while LLMs can effectively sometimes surpass human regulators in EER, such advantage seems to be limited to providing actionable tactics within comforting messages. However, our results also suggest overall advantages seem to be context and outcome dependent. I will discuss the implications for theories of interpersonal emotion regulation and AI-human interaction, highlighting the critical role of actionable regulatory tactics for successful interpersonal emotion regulation.