A 12-month longitudinal study of adults in four English-speaking countries found that lonely people may be more likely to seek companionship through chatbots. However, such use can worsen feelings of loneliness over time. The paper was published in psychological science.
The recent development of large language models, AI systems that can communicate with users in natural human language, has ushered in a new era for humanity. It can be said that a revolution has begun, both in the way business is done and in the way people live. Immediately after its release at the end of 2022, ChatGPT gained millions of users. It is estimated that in 2026, over 1 billion people will be using large-scale language models and generative AI tools.
In this short period of time, many people have started using AI chatbots for conversations. Some researchers have suggested that AI companionship could become a scalable tool to combat the loneliness epidemic. But some warn that AI companions could do more harm than good. The main reason for this is that AI is unable to truly experience emotions or bond with other people. It can only give the illusion of understanding, empathy, compassion, and love.
Additionally, AIs lack a true inner self, making them incapable of the reciprocal self-disclosure that is essential to building long-term, rewarding relationships. As a result, forming bonds with AI companions or AI romantic partners may ultimately leave vulnerable individuals unsatisfied and lead to worsening mental health over time.
Study authors Dunnigan Falk and Elizabeth Dunn point out that people may rely on AI companions because they are always available (unlike human partners or friends) and can simulate human companionship in a way that feels compelling in the moment. But researchers theorize that these simple but shallow interactions with AI may crowd out more valuable interactions with real humans in people’s lives.
To investigate this, they conducted a 12-month longitudinal study investigating the relationship between loneliness and the use of chatbots for social purposes. Study participants included 2,149 people from the United Kingdom (50%), United States (28%), Canada (14%), and Australia (8%) who completed at least one survey across four data collection waves. Of these participants, 979 completed all four surveys and 466 completed three surveys. The average age of participants was 40 years, and 49% were male.
The survey asked participants how often they used chatbots for social purposes in the past four months, such as seeking advice on life decisions, having regular social conversations, and seeking social companionship. The study also measured “emotional isolation” by asking participants how emotionally isolated they felt from others during the same period.
Additionally, the researchers tracked broader indicators of overall “social connectedness” and asked participants whether they had experienced any major life events, such as a job transfer, a breakup, the beginning of a stable romantic relationship, or becoming parents, to control for external social stressors.
Results showed that approximately 26% to 30% of participants across waves reported using chatbots for social purposes. On average, the frequency with which participants used chatbots for these purposes did not change significantly over the data collection period.
Importantly, people who reported feeling emotionally isolated at one point were more likely to report using chatbots for social purposes more four months later. This confirmed the study authors’ prediction that people would seek out chatbots when they felt lonely. However, after increasing chatbot use, participants also tended to report more psychological isolation at the following time points: On the other hand, experiencing a major life event was not associated with increased subsequent use of chatbots.
When the study authors measured overall social connectedness more broadly, the results showed that people who felt less socially connected tended to increase their use of social chatbots after four months. However, in this analysis, increased chatbot use did not lead to further declines in social connectedness at later time points. (Experiencing a breakup tended to decrease feelings of social connection four months later, but it was not associated with increased use of chatbots. No other types of life events were associated with changes in social connection.)
“Taken together, these findings provide initial evidence that feelings of loneliness may prompt people to seek companionship through chatbots, but that such use may exacerbate feelings of loneliness over time. However, given the exploratory nature of the analysis, we urge caution in drawing strong conclusions,” the study authors concluded.
These findings contribute new evidence on how people use AI chatbots to meet their psychological needs. However, it should be noted that the observational design of this study does not allow definitive causal inferences to be drawn from the results. Furthermore, all findings were based on self-reporting, leaving room for reporting bias to influence the results.
The paper, “How does using AI for peer relationships predict loneliness and vice versa?” is written by Dunnigan Falk and Elizabeth Dunn.
