Artificial Intelligence (AI) chatbots are increasingly being explored as sources of informal emotional support, with emerging evidence suggesting that venting to these systems can reduce negative affect. Yet, it remains unclear whether such benefits depend on the responder's perceived identity. Given that emotional relief from venting often hinges on perceived authenticity and emotional validation, this study investigates whether the emotional well-being benefits of venting differ when users believe they are interacting with an AI chatbot versus a human, even when responses are content-matched. In a pre-registered experiment ( N = 279), participants were randomly assigned to either an AI-assisted venting condition or a perceived human-assisted venting condition. Importantly, participants in both conditions received similar responses generated by an AI chatbot. In the perceived human-assisted venting condition, however, these responses were slightly edited and presented in a context designed to enhance their credibility as coming from a real person. Results indicated that venting improved some emotional well-being outcomes in both conditions, including reductions in stress and loneliness, and increases in perceived social support. However, the magnitude of improvement in emotional well-being outcomes was similar across the AI-assisted venting and perceived human-assisted venting conditions. These findings suggest that AI chatbots may deliver emotional benefits comparable to those provided by a perceived human responder, even when users are aware they are interacting with an artificial agent.
Hu et al. (Wed,) studied this question.