Generative artificial intelligence-assisted visual art therapy improves cognition, anxiety, and depression in community-dwelling older adults.

OBJECTIVE: To examine whether generative artificial intelligence-assisted visual art therapy improves cognitive function and alleviates anxiety and depression in community-dwelling older adults with mild cognitive impairment. METHODS: A quasi-experimental, cluster-gated design was employed in Shengli Street Community, Enshi City (September 2025-April 2026). Eighty eligible older adults were enrolled and assigned by administrative grid to an intervention group (n = 40) or a control group (n = 40). The control group continued routine community activities; the intervention group additionally received a 12-week generative artificial intelligence-assisted visual art therapy program. The primary outcome was change in Montreal Cognitive Assessment (MoCA) total score; the Hospital Anxiety and Depression Scale (HADS) served as the secondary outcome. Assessments occurred at baseline, 1 week post-intervention, and at three-month follow-up. Cohen's d effect sizes are reported. RESULTS: Seventy-two participants completed the study. At post-intervention, the intervention group scored significantly higher on the MoCA (d = 1.36, p < 0.001) and lower on HADS-Anxiety and HADS-Depression (d = 0.73-0.76, both p < 0.01) relative to controls. Gains were largely maintained at 3 months (MoCA d = 1.08). Fidelity reached 93.5%; no serious adverse events occurred. CONCLUSION: An AI-enabled, four-phase visual art therapy program can improve cognition and reduce anxiety and depression in community-dwelling older adults with mild cognitive impairment, with effects partially sustained at 3 months. The observed benefits reflect the synergistic effects of the multi-component intervention as a whole, rather than the isolated contribution of any single element.