Turning to Chatbots when Lonely May Exacerbate Feelings of Loneliness, Study Finds

Why It Matters

The findings suggest that relying on AI chatbots for companionship could worsen loneliness, raising concerns for mental‑health outcomes and prompting tech firms to reconsider how these tools are positioned and integrated.

Key Takeaways

• 26‑30% of adults used chatbots for social purposes each wave
• Lonelier participants increased chatbot use four months later
• Increased chatbot use predicted higher emotional isolation later
• Major life events did not drive chatbot usage
• Observational design limits causal conclusions about AI companionship

Pulse Analysis

The rapid diffusion of large‑language‑model chatbots has reshaped daily interaction patterns. Since ChatGPT’s launch in late 2022, estimates suggest more than one billion users worldwide now engage with generative AI tools, many seeking not only information but also companionship. This surge coincides with a growing loneliness epidemic, prompting researchers to explore whether AI can serve as a scalable remedy. The new longitudinal study published in Psychological Science adds empirical weight to the debate, examining how emotional isolation drives chatbot use and how that use, in turn, influences feelings of loneliness.

The researchers followed 2,149 adults from the United States, United Kingdom, Canada and Australia over twelve months, collecting four waves of self‑reported data on chatbot interactions and emotional isolation. Roughly a quarter to a third of participants regularly turned to chatbots for social conversation, and those who felt more isolated at one wave were statistically more likely to increase chatbot use four months later. Crucially, higher chatbot engagement subsequently predicted a rise in emotional isolation, suggesting a feedback loop where shallow AI interactions may displace richer human connections. Broader social‑connection scores showed a similar pattern without a reciprocal decline.

These findings carry weight for developers, mental‑health practitioners, and policymakers. While AI companions can offer immediate, judgment‑free dialogue, the data warn that reliance on them may exacerbate the very loneliness they aim to alleviate, especially for vulnerable users. Companies should consider integrating prompts that encourage offline social activity or referrals to professional support. Clinicians might screen for excessive chatbot use as a marker of worsening isolation. Future research must move beyond self‑report surveys to experimental designs that can untangle causality and guide responsible AI deployment.