De Freitas et al.: adult AI companions can reduce momentary loneliness

De Freitas, Oğuz-Uğuralp, Uğuralp & Puntoni, Journal of Consumer Research (online June 2025; volume April 2026); full author paper. This is the positive original adult counterpoint to Li’s two-week human comparison and the adolescent question.

Several studies distinguish immediate relief after a companion exchange from population-level development. In the preregistered study 5, 2,228 adult online participants (mean age 40.7) were assigned a single AI-companion exchange, journaling or a limited-functionality bot; post-exchange loneliness was lower in the companion condition (21.84 versus 33.65 for journal and 30.71 control on authors’ scale, both p<.001). In a week of repeated chat, the authors report after-chat momentary loneliness improvement; that is not evidence six-month changes in offline relationships. Study 2’s human exchange comparison on immediate loneliness likewise does not contradict Li et al.’s two-week between-arm advantage for human-peer contact; populations, bots, comparators and endpoints differ.

A skeptic of blanket under-18 exclusion can reasonably ask whether a low-risk bounded companion benefits an isolated young person without available peers; no current adult study settles this for minors. An experimental minor design would benchmark not merely diary/no support but actually available human support, measure acute feeling heard, longer-run peer contact and referral completion, and report adverse tails and dropout.