Implicit and Explicit Partner Preferences of Individuals in Human-Chatbot Relationships

Humans have long been considered the gold standard of social interaction partners. Recently, however, socially capable technologies, including social chatbots, have become increasingly able to engage users in emotionally, romantically, and even sexually meaningful ways. For individuals who are currently in romantic relationships with chatbots, it is therefore an open empirical question to what extent human partners remain the preferred partner, and how concerns about “replacement” should be understood in this specific population. The present study addresses this question by recruiting individuals who are currently in a romantic or emotionally intimate relationship with a chatbot and assessing both their implicit attitudes and explicit partner preferences. Participants complete an Implicit Association Test (IAT) contrasting human–human and human–chatbot romantic relationships on a good–bad evaluative dimension. They then explicitly indicate whether they would prefer to date a human or a chatbot partner. To better understand these preferences, participants also answer brief open-ended questions and factual items about their relationship history (length of the current chatbot relationship, whether they have had previous human–human relationships, who ended the most recent human relationship, and whether they experienced psychological or physical trauma in a human relationship). By combining implicit measures with explicit reports, the study aims to provide a clearer picture of whether, for people currently in chatbot relationships, human–human relationships continue to be the partners of choice or whether chatbot partners are preferred, and how prior human relationship experiences may relate to these evaluations. The goal is not to generalize to society at large, but to clarify partner preferences within this population, thereby informing theory and public discussion with evidence rather than alarmist assumptions.