AI Companions Reduce Loneliness

Julian De Freitas,Ahmet K Uguralp,Zeliha O Uguralp,Puntoni Stefano

2024-07-09

Abstract:Chatbots are now able to engage in sophisticated conversations with consumers in the domain of relationships, providing a potential coping solution to widescale societal loneliness. Behavioral research provides little insight into whether these applications are effective at alleviating loneliness. We address this question by focusing on AI companions applications designed to provide consumers with synthetic interaction partners. Studies 1 and 2 find suggestive evidence that consumers use AI companions to alleviate loneliness, by employing a novel methodology for fine tuning large language models to detect loneliness in conversations and reviews. Study 3 finds that AI companions successfully alleviate loneliness on par only with interacting with another person, and more than other activities such watching YouTube videos. Moreover, consumers underestimate the degree to which AI companions improve their loneliness. Study 4 uses a longitudinal design and finds that an AI companion consistently reduces loneliness over the course of a week. Study 5 provides evidence that both the chatbots' performance and, especially, whether it makes users feel heard, explain reductions in loneliness. Study 6 provides an additional robustness check for the loneliness alleviating benefits of AI companions.

Computers and Society,Artificial Intelligence

What problem does this paper attempt to address?

The problem this paper attempts to address is whether AI Companions can effectively alleviate users' feelings of loneliness. Specifically, the study explores this issue through the following aspects:

1. **Detecting Loneliness**: Researchers used a novel approach by fine-tuning large language models (LLMs) to identify loneliness exhibited by users during interactions with AI companions. This method is more accurate than traditional dictionary-based methods.

2. **Evaluating Effectiveness**: Through multiple studies (including cross-sectional and longitudinal studies), the changes in users' loneliness before and after interacting with AI companions were assessed and compared with other methods of alleviating loneliness (such as interacting with real people, watching videos, etc.).

3. **Exploring Mechanisms**: The study also investigated the factors that lead to AI companions being able to reduce loneliness, with a particular focus on the feeling of "being heard" and the performance of chatbots.

4. **Predictive Bias**: The accuracy of users' predictions regarding the effectiveness of AI companions in reducing loneliness was examined, revealing that users often underestimate the actual effectiveness of these applications.

In summary, this paper aims to empirically verify whether AI companions can effectively reduce loneliness and further explore the underlying mechanisms.