(PS8-1) Ai’ll Be There for You: Couples, AI, and Relationship Maintenance

The development of new technologically assisted communication and, more recently Artificial Intelligence (AI), has given couples novel communication mediums to support  relationship maintenance. Outcomes for couples, when technology is used, are quite mixed. Some studies show that technologically aided communication can foster a sense of connection and that there is no difference from in-person communication (Kilpatrick, 2007; Perry & Werner-Wilson, 2011). However, other studies (e.g., McDaniel & Coyne, 2016) show strong deleterious effects of “technoference” on relationship satisfaction. This is alarming as there is also an increasing trend for the utilization of various forms of technology in couples-based treatments for distress (e.g., Doss & Hatch, 2022). 

 Studies have suggested that AI utilization tends to be perceived more negatively by recipients (e.g., Hohenstein et al, 2023; Liu et al., 2023), but there has not been a study on the perceptions of AI within couples relationships. Based on mixed research on the impact of technology on couples communication, paired with the advancement of AI technology, research is necessary to gauge how these new technologies have influenced or changed couples communication and relationship outcomes. 

Data for the current study are being collected from undergraduate students in romantic relationships at a midwestern university. Participants were randomly assigned to one of three experimental groups (AI-assisted communication, human/friend-assisted communication, or unassisted communication). Within these experimental groups, there were three different relational tasks or hypothetical scenarios that were presented to participants (e.g., tired after a long work day or family birthday plannin). Participants were asked to create a message to describe this scenario to their partner and were provided with a stem response that was based on their experimental group randomization. They were then provided with a stem response (e.g., “Hey, sorry to hear about your burnout”) and a description of how their partner generated the response (i.e., experimental condition). After each task, participants were asked to rate their affect, perception of effort, perceived response appropriateness, and response satisfaction. It is expected that participants in the AI condition will have more negative perceptions of response than in the other conditions. However, familiarity and relationship satisfaction will be explored as potential moderators of this effect. Given the innovative potential of AI in couples-based treatments, it is imperative that similar research provide basic information on partner responses to various types of technology assisted communication as a form of relationship maintenance communication.