In a PRISMA-guided systematic review and meta-analysis of randomized controlled trials, AI-based socially assistive robots (AI-SARs) in Alzheimer’s disease and related dementias (ADRD) were associated with a small but statistically significant reduction in depression (SMD -0.30; 95% CI -0.56 to -0.03; p=.027) and agitation (SMD -0.26; 95% CI -0.49 to -0.03; p=.027). Effects on cognition (SMD -0.07; p=.555) and anxiety (SMD 1.12; p=.440) were not significant.

Why It Matters To Your Practice

• Neuropsychiatric symptoms drive distress, caregiver burden, and medication use in ADRD — so even small nonpharmacologic gains in mood or agitation can be clinically meaningful.

• Robotic “companions” are moving from pilots to procurement decisions (senior living, home health), and clinicians will increasingly be asked whether they work and for whom.

Clinical Implications

• Position AI-SARs as an adjunct to established nonpharmacologic strategies (activity scheduling, behavioral/environmental approaches), not a replacement for clinical assessment or caregiver training.

• Set expectations: evidence supports modest symptom improvements (depression/agitation), but not cognition or anxiety in current RCTs.

• When families ask: discuss practical selection criteria (ease of use, supervision needs, infection control, sensory limitations, fall risk around devices) and how response will be monitored over time.

Insights

• Magnitude: depression improvement was small (SMD -0.30) — roughly a “small effect size,” which may still matter when options are limited or medication risks are high.

• Signal appears symptom-specific: benefits clustered in affect/agitation rather than cognitive outcomes, consistent with a mechanism of social engagement/structured interaction.

• Evidence gaps remain: heterogeneity in robot type, intervention “dose,” and outcome measures likely contributes to the inconclusive findings for anxiety and cognition.

The Bottom Line

• AI social robots in ADRD RCTs show small reductions in depression and agitation, with no clear benefit for cognition or anxiety.

• Clinically, treat AI-SARs as a potentially useful adjunct for select patients — while advocating for larger, standardized trials before broad claims about cognitive or anxiety benefits.