AI Phone Care Calls Show Promise for Dementia Patients’ Memory and Depression
A peer-reviewed study published in Scientific Reports found that AI care calls twice weekly significantly reduced depression scores.
It also improved memory function in community-dwelling individuals with dementia over a 31-week intervention.
Moreover, the findings align with AI care calls, dementia memory improvement, AI depression treatment, elderly cognitive health.
The 80-participant study offers real-world evidence that AI tools can aid elderly cognitive health.
Overview
Artificial intelligence care calls represent an emerging form of remote health support.
AI systems conduct structured phone conversations with older adults to monitor well-being, provide emotional engagement, and encourage cognitive activity.
Additionally, the most rigorous evidence to date comes from a study published in Scientific Reports on July 25, 2025.
It was led by researchers at Hanyang University Hospital in South Korea.
The study involved 80 community-dwelling individuals with dementia. They averaged 79.9 years of age and received AI care calls twice weekly for 31 weeks. The AI system used was CLOVA CareCall, a large language model-based conversational agent developed by Naver. It is designed specifically for voice-based phone interactions with older adults.
Cognitive function was measured using standardized assessments of attention, memory, and language. Depression was measured using the Geriatric Depression Scale (GDS). Both measures were taken before and after the intervention period.
Why This Matters
The study found statistically significant improvements in two key areas:
Depression reduction: GDS scores decreased from a median of 8.5 to 6.0 after the intervention. At baseline, 42.5% of participants (34 individuals) had major depressive symptoms (GDS ≥ 10). After the intervention, 44.1% of those with major depression improved to minor or non-depressed status. Additionally, 42.9% of those with minor depression improved to non-depressed status. This distributional shift was statistically significant (Stuart-Maxwell test, P = 0.002).
Memory improvement: Memory scores increased from a median of 3.0 to 4.0 (p < 0.001). Attention and language scores showed no changes. Thus, it aligns with AI care calls, dementia memory improvement, AI depression treatment, elderly cognitive health.
The results are notable because the intervention required no special equipment, training, or in-person visits. Additionally, participants simply answered a phone call — an activity many older adults already understand and accept.
This is critical because the number of people with dementia worldwide will rise. It will grow from 57.4 million in 2019 to 152.8 million by 2050, increasing care burdens.
Key Takeaways
- AI care calls reduced depression significantly: A structured AI phone conversation program reduced depressive symptoms in dementia patients by an average of 2.5 points on the GDS scale. Nearly half of participants with major depression improved to a less severe category.
- Memory scores improved: The median memory assessment score increased by one full point, the only cognitive domain showing statistically significant change.
- Sex and education matter: Female participants showed greater memory improvement and maintained attention scores, while male participants experienced attention decline with no memory gain. Higher education levels were associated with a protective effect against worsening depression and attention scores.
- The AI detected and responded to emotional cues: A follow-up study published in BMC Geriatrics (January 28, 2026) examined 2,896 older adults using the same AI call system. While the AI identified depressive cues in conversations, it also misinterpreted or missed them in some cases highlighting that current AI lacks full contextual understanding of human emotional expression.
- No special hardware required: Unlike AI care robots or video-based systems, phone-based AI care calls work on any telephone, making them the most accessible form of remote AI health support for older adults in both urban and rural settings.
Practical Considerations
Current limitations: The study was conducted in South Korea using a Korean-language AI model (HyperCLOVA x). Participants were primarily female (73.8%). Additionally, generalizability to other populations and languages requires further testing. Moreover, the AI system’s difficulty in detecting nuanced emotional cues has been documented in the 2026 BMC Geriatrics study. Thus, these systems cannot replace professional mental health evaluation or diagnosis.
What families can do: AI care call technology is still emerging and not yet widely available in most markets. However, families should consider any AI-powered health monitoring tools that may become available in memory care facilities. The evidence suggests these tools work best as supplements to human care, not replacements. Regular phone check-ins from family members or volunteers remain equally important.
Additionally, the study notes caregiver burden and cites research showing AI care systems could prevent solitary deaths in elders. This is a significant concern for families caring for relatives with dementia at risk of undetected health crises.
Ethical considerations: Researchers note that AI care calls offer scalability.
Additionally, they must be designed to respect residents’ autonomy and ensure data privacy.
Moreover, the risk of over-reliance on technology as a substitute for human connection remains a valid concern.
Sources
- Kang, S., Lee, Y., Kim, N., et al. “Beneficial effect of artificial intelligence care call on memory and depression in community dwelling individuals with dementia.” Scientific Reports, 15, 27116 (2025). https://doi.org/10.1038/s41598-025-12895-7
- Kim, D., Jang, R., Jang, S., Park, S. “When a chatbot asks ‘How are you?’: A cross-sectional study of AI call conversations and depressive symptom detection among older adults in rural South Korea.” BMC Geriatrics, 26, 247 (2026). https://doi.org/10.1186/s12877-026-07038-0
- “The Promise of AI Companions in Memory Care.” Provider Magazine, American Health Care Association, February 18, 2025. https://www.providermagazine.com/Articles/Pages/The-Promise-of-AI-Companions-in-Memory-Care.aspx
Keywords: AI care calls, dementia memory improvement, AI depression treatment, elderly cognitive health