Does AI Reduce Loneliness? What the Research Says
Four recent studies: what they show—and what they don’t. A sober look at the evidence.
The question sounds simple. The answer isn't.
Can artificial intelligence really reduce loneliness among older adults—or is that a promise that sounds better than it delivers? The body of research on this topic has grown significantly over the past two years. This article summarizes the findings of the most robust studies to date—and notes what they do not answer.
“Thequestion of whether AI reduces loneliness sounds simple. The answer is not.”
Study 1: Harvard and Wharton, measurable reduction in controlled experiments
De Freitas et al. (2024) published a series of experiments in which participants interacted either with an AI companion app or with no intervention at all. The result: AI companionship measurably reduced participants’ subjective feelings of loneliness —with effect sizes comparable to those of brief human social interactions.
The authors emphasize that the effect is not based on deception. Participants knew they were interacting with AI—and yet still found the connection meaningful. This contradicts the commonly held assumption that connecting with AI requires ignorance of the nature of the conversation partner.
Limitation: The studies were conducted in a laboratory setting, not in the everyday lives of older adults. The long-term effects have not been documented.
Study 2: Meta-analysis of 19 studies, significant effects, but context-dependent
A meta-analysis published in 2025 (PMC) synthesized 19 studies involving a total of 1,083 older participants that had examined interactions with social robots and AI companions. The overall finding: the interventions statistically significantly reduced loneliness.
Particularly noteworthy: The effect was more pronounced among people in long-term care facilities than among those living independently. This suggests that AI-assisted companionship is most beneficial where access to human contact is most limited—not as a substitute for existing social structures, but as a support where such structures are virtually nonexistent.
Limitation: The quality of the studies varies considerably. Many studies lack control groups, have short observation periods, and small sample sizes. The authors themselves describe the overall evidence as promising but not yet conclusive.
Study 3: Japan, 14,721 adults – Well-being increases, but only under certain conditions
The largest study to date on this topic (2024/2025, Japan, N = 14,721) examined the relationship between the use of AI companions and subjective well-being across three dimensions: life satisfaction, happiness, and a sense of meaning. The surprising finding: AI companionship increased well-being more significantly among people with strong social networks —not among those who were most isolated.
This contradicts a simple causal relationship. The authors’ interpretation: AI-assisted support may act as an enhancer of social skills and the willingness to connect—not as a substitute. Those who are already socially integrated use AI as a supplementary means of connection. Those who are deeply isolated benefit less, possibly because chronic loneliness impairs the ability to connect in the first place.
Limitation: Cross-sectional study; causality cannot be established. Cultural characteristics specific to Japan may not be directly applicable.
“AI-assisted supportmay serve to enhance people’s willingness to connect socially—not as a substitute for it.”
Study 4: The IQWiG and German Data, Positive Indications, Weak Evidence
In 2022, the Institute for Quality and Efficiency in Health Care (IQWiG) published a comprehensive HTA report on measures to combat social isolation among older adults. Digital support was one area covered. The conclusion: There is some evidence of positive effects, but the overall evidence is rated as low.
This is neither a green light nor a rejection. It is simply a statement of the current state of research: The studies available so far are too methodologically diverse, too short-term, and too small in scale to allow for definitive conclusions. That is changing right now—but it takes time.
What the research as a whole shows
Taken together, the four studies paint a picture that warrants neither jubilation nor skepticism. What can be said with increasing certainty is:
AI companionship can reduce loneliness. The effects are measurable, consistent across multiple studies, and do not depend on users’ lack of awareness regarding the nature of the interaction.
Context is key. The impact is greater in situations where human contact is structurally limited—and may be less pronounced among people who are already deeply isolated.
Design is crucial. Studies examining apps designed to promote well-being yield different results than those testing products designed to foster addiction.
The long-term effects have largely not been studied. What happens after six months or a year—whether the effects remain stable or whether unexpected side effects arise—is not yet fully understood.
This means that the evidence is sufficient to consider AI-assisted therapy a legitimate tool. It is not sufficient to resolve all outstanding questions. Anyone who claims the opposite—in either direction—is going beyond what the research can currently support.
References
De Freitas, J., Uğuralp, A.K., Uğuralp, Z., & Puntoni, S. (2024). AI Companions Reduce Loneliness. Harvard Business School Working Paper No. 24-078.
PMC. (2025). Wired for companionship: a meta-analysis on social robots filling the void of loneliness in later life. 19 studies, N = 1,083. Robust Variance Estimation.
ScienceDirect. (2025). AI companions and subjective well-being: Moderation by social connectedness and loneliness. N = 14,721 Japanese adults. December 2024 / January 2025.
IQWiG. (2022). Social Isolation and Loneliness in Older Adults: What Measures Can Prevent or Counteract Social Isolation? HTA Report No. 1459.
