Skip to content

Artificial agents may ease loneliness in older adults

Pooled findings favored artificial agents for loneliness and depression; quality of life and happiness showed no statistically significant improvement.

By 100HP editorialAbstract-based explanation checked

Based on the published abstract. The full paper may contain additional methods, results and limitations.

The 30-second takeaway

Older adults receiving artificial agent interventions had lower loneliness scores in the pooled analysis. Depression scores also improved, but quality of life and happiness did not show statistically significant effects. The results suggest a possible benefit for specific symptoms, while leaving uncertainty about broader wellbeing, how long any benefit lasts, and which approach works best.

Keep in mind

Results varied across studies, and the abstract lacks follow-up durations and trial details needed to assess subgroup comparisons. Its statements about longer and shorter sessions are difficult to reconcile.

THE NUMBERS, WITH CONTEXT

What researchers found

-0.69 (95% CI -0.89 to -0.40)

Loneliness scores: standardized mean difference

The group difference is measured relative to score variability, in standard deviations, not scale points or percentages. Negative favors agents over no agent intervention; pooled assessment timing was not reported.

Arch Gerontol Geriatr, 2026 · Original source ↓

Separate wellbeing outcomes

Depression scores improved in the pooled analysis. Quality of life and happiness did not show statistically significant effects; those findings leave uncertainty about benefit and do not demonstrate that effects are absent.

Agent comparisons need care

Subgroup analyses favored social robots over app or web agents, and animal-shaped agents over humanoid ones. These patterns do not establish which agent would work best for an individual older adult.

CHECK THE ORIGINAL

The original publication

The effectiveness of artificial agents on loneliness reduction in older adults: a systematic review and meta-analysis.

Fu Z, Yan M, Han X et al.
Arch Gerontol Geriatr · 2026

PubMed ID
41124893
Record checked

AI-assisted research and writing. This explains one selected publication; it is not a complete review of everything known. Our approach.