Interactive Learning Environments

AI interaction linked to less digital loneliness in older adults

A four-week randomized trial of 90 older adults found a greater decline in digital loneliness with AI-assisted interaction than with conventional interaction.

Paper by WeiCheng Pan, WanQing Zhang, Hong Chen, 2 more authorsHesen Li · Shenyang Aerospace University
Older adults gather around a tablet amid many speech bubbles, while gray figures stand farther apart around another tablet with fewer bubbles.
Randomized trial in 90 older adults, four-week, against conventional interaction control groupInterventions

210 participants

Limit: The loneliness reduction was observed over only a four-week intervention

Edited by Boris Djordjevic · AgingResearch.News

Follow

In a four-week randomized trial, 90 older adults took part in artificial intelligence (AI)-assisted interaction or conventional interaction. The AI-assisted group showed a greater reduction in digital loneliness over the study period.

A second experiment involved an independent sample of 210 participants. After accounting for initial digital loneliness, statistical analyses identified perceived social support—how supported participants felt—as an indirect link between AI-assisted interaction and lower loneliness after the intervention. The association between AI-assisted interaction and perceived support also varied with participants’ acceptance of technology. Evidence for a similar pathway through participants’ assessments of emotional fulfillment remained insufficient. Together, the findings suggested that the association with reduced digital loneliness was primarily related to perceived social support.

Why it matters

For older adults using digital learning systems, the findings suggest that designers should consider both acceptance of the technology and its capacity to offer social support.

Caveats

The statistical pathway analysis does not establish that feeling supported caused the reduction in loneliness. The abstract did not report participants’ age ranges, the size of the reductions or harms, and did not specify whether the second experiment was randomized.