Are AI lovers replacing real romantic partners? Surprising findings from new research

A new study published in the Archives of Sexual Behavior offers insight into how romantic relationships with virtual agents—artificially intelligent characters designed to simulate human interaction—might influence people’s intentions to marry in real life. The research found that these virtual relationships can affect marriage intention in both negative and positive ways, depending on the psychological mechanisms involved. Feelings of connection and authenticity in these digital relationships were linked to lower interest in marriage, while emotional satisfaction and a sense of personal competence were associated with more favorable attitudes toward marrying a real partner.

The study was motivated by growing concerns about the declining rates of marriage and family formation, especially in East Asian cultures like China where social norms strongly emphasize marriage and filial duty. As technology evolves, a growing number of people are turning to virtual agents—non-human but human-like characters found in video games and apps—to fulfill emotional and romantic needs. These relationships are particularly common in romantic simulation games, where users engage in conversations and storylines with digital characters.

Some researchers worry this trend may reduce people’s motivation to pursue real-life romantic relationships, while others suggest these digital experiences could help people improve social skills and build confidence for real-world intimacy. The current study aimed to examine both possibilities and uncover the psychological processes that mediate the relationship between digital love and real-life marriage intentions.

To explore these questions, the researchers conducted an online survey between December 2023 and February 2024. The sample included 503 Chinese participants—182 men and 321 women—who had engaged in romantic relationships with virtual agents in the past year. Participants were recruited through snowball sampling, beginning with players of well-known romantic games like “Light and Night,” “Mr. Love: Queen’s Choice,” and “VR Kanojo.” Respondents were asked about the amount of time they spent with their virtual partners, their emotional experiences, and their attitudes toward real-life marriage.

The survey measured several key concepts, including satisfaction of relatedness (how emotionally fulfilling the virtual relationship felt), relationship authenticity (how “real” the connection seemed), and desire for a real-life relationship with the virtual agent. It also included assessments of participants’ emotional responses, feelings of immersion and competence in the relationship, and standard psychological measures of marriage attitudes, perceived social norms, and sense of control over marriage-related decisions. Participants also reported their overall intention to marry a real person.

Analysis showed that time spent with virtual agents was positively associated with feelings of emotional closeness, perceived authenticity of the relationship, and experiences of enjoyment, immersion, and confidence. However, those same feelings were also linked with reduced interest in marriage, especially among male participants. One explanation is that when virtual relationships satisfy people’s need for emotional connection, they may feel less inclined to seek out similar relationships in real life. This pattern was particularly pronounced among men, possibly due to differing motivations or mating strategies.

Another mechanism at play was romantic anthropomorphism—the tendency to perceive virtual agents as emotionally genuine partners. Participants who felt a strong emotional bond with their digital companions were more likely to say they wished their virtual agent could be real. These individuals were also less inclined to want to marry a real person, possibly because their idealized views of virtual romance led them to judge real-life relationships less favorably.

At the same time, the study found that virtual relationships could boost people’s confidence and enhance their attitudes toward marriage. Participants who felt happy and competent during interactions with their virtual agents were more likely to express positive views about marriage and reported a stronger sense of control over their ability to achieve it. These findings suggest that virtual relationships may provide emotional benefits that carry over into real-world attitudes, potentially increasing marriage intention through improved mood and a greater sense of capability.

Interestingly, the researchers had expected that immersion in virtual relationships might disconnect users from real-world social norms, leading to lower marriage intention. But the data told a different story. Participants who reported higher levels of immersion were actually more likely to endorse traditional marriage norms. This finding implies that virtual romantic experiences may reinforce rather than undermine cultural expectations—perhaps because the narratives and character dynamics in romantic games often reflect real-life social scripts.

Gender differences emerged across several pathways. Men were more strongly affected by feelings of relatedness, immersion, and competence in virtual relationships than women. These differences may reflect both cultural norms and underlying biological tendencies. Previous research has suggested that men may be more prone to forming emotional attachments to non-human companions, especially when those relationships offer a sense of control, validation, or idealized connection.

The study has several practical implications. On one hand, it raises concerns that people who find emotional fulfillment in virtual relationships may become less motivated to seek out real-life partners. On the other hand, it also suggests that virtual experiences could be designed to promote healthy attitudes toward marriage. Developers of romantic games could embed themes that highlight the value of real-life commitment, while mental health professionals might help individuals distinguish between digital escapism and genuine relational needs.

Like all research, this study has limitations. It relied on a cross-sectional survey design, which means it cannot determine cause and effect. It’s possible that people with lower interest in marriage are more likely to engage in virtual relationships, rather than the other way around. Future studies using longitudinal or experimental designs are needed to clarify the direction of these relationships. In addition, because the sample was drawn from a single country and relied on a specific recruitment method, the findings may not generalize to other cultural contexts or populations.

The study, “Romantic Relationships with Virtual Agents and People’s Marriage Intention in Real Life: An Exploration of the Mediation Mechanisms,” was authored by Jia‑Lin Zhao, Ru Jia, John Shields, Yu‑Jia Wu, and Wei‑Wei Huang.