KOR-004

Hyodol AI Care Robot for Elderly Living Alone (South Korea)

Download PDF
Korea, Rep. East Asia & Pacific High income Operational Deployment (Limited Rollout) Confirmed

Local government welfare centres (e.g. Guro District, Seoul); Ministry of Health and Welfare and National Health Insurance Service (national long-term care pilot from September 2024); Hyodol Co., Ltd. (technology provider). (Rest of World 2025; Aju Press 2024)

At a Glance

What it does Perception and extraction from unstructured inputs — User communication and interaction
Who runs it Local government welfare centres (e.g. Guro District, Seoul); Ministry of Health and Welfare and National Health Insurance Service (national long-term care pilot from September 2024); Hyodol Co., Ltd. (technology provider). (Rest of World 2025; Aju Press 2024)
Programme Hyodol AI Care Robot Distribution Programme for Elderly Living Alone
Confidence Confirmed
Deployment Status Operational Deployment (Limited Rollout)
Key Risks Market, sovereignty and industry structure risks
Key Outcomes Jung et al.
Source Quality 4 sources — News article / media, Academic journal article

Hyodol is an AI-enabled socially assistive companion robot deployed at scale through South Korea's government welfare programmes to provide care, companionship, and health monitoring to elderly people living alone. The name Hyodol is a compound of 'hyo' (the Confucian value of filial piety) and 'doll', reflecting the robot's design as a caring grandchild figure for isolated older adults. Developed by Hyodol Co., Ltd. (a South Korean technology company), the robot takes the form of a soft, doll-like figure resembling a young child, with anime-style eyes and neon-red cheek lights. Despite its toy-like appearance, Hyodol incorporates advanced AI and sensor technologies to deliver a range of health and welfare functions.

The robot's core AI capability is a ChatGPT-powered conversational chatbot that enables natural-language voice interaction with users in Korean. This allows elderly users to engage in free-form conversation with the robot, which responds in a chirpy, grandchild-like persona designed to foster emotional connection and reduce feelings of isolation. Beyond conversational companionship, Hyodol provides structured daily health check-ins that assess users' mood, pain levels, meal consumption, medication adherence, and sleep quality. These check-ins are recorded and transmitted to social workers and family members via a companion smartphone application, enabling remote monitoring of the elderly person's wellbeing.

Hyodol's hardware includes multiple sensor systems. An infrared neck sensor detects user movement and presence, and the system is configured to flag alerts if no movement is detected for 24 hours, triggering welfare checks by social workers or family members. A microphone embedded in the chest records daily verbal responses during health check-ins, and a Microsoft AI-enabled voice analysis programme assesses these recordings to evaluate the user's emotional state and mood over time. Touch-sensitive sensors are distributed across the robot's head, hands, ears, and body, responding to physical interaction and reinforcing the companion relationship. The robot also incorporates fall detection capabilities and can send emergency alerts to designated caregivers.

Functionally, Hyodol serves as a medication reminder system, prompting users at scheduled times to take prescribed medications. It provides activity prompts encouraging physical movement such as walks, offers cognitive stimulation through dementia-prevention quizzes, and delivers content including singing, guided exercise routines, and religious materials tailored to user preferences. The robot is designed to support behavioural activation principles for mental health, particularly targeting subclinical and mild depressive symptoms common among isolated elderly populations.

The deployment of Hyodol has been substantially driven by government welfare programmes at both national and local levels. South Korea's government has pursued a decade-long 'robotizing' policy for eldercare, beginning with the Ministry of Knowledge Economy's 'Robot Future Strategy' announced in 2012, which allocated funding to local municipalities and welfare institutions to purchase robotic products for socially disadvantaged populations. Hyodol has been a primary beneficiary of these programmes. The Guro District of Seoul began distributing Hyodol robots in 2019, investing 200 million won (approximately USD 143,867) with support from the federal industrial technology ministry, at a unit cost of approximately 1.6 million won (USD 1,150) per robot. As of reporting in 2025, over 12,000 Hyodol units have been deployed to the homes of solitary elderly individuals across South Korea, with over 1,300 units distributed in the South Jeolla province alone.

In September 2024, Hyodol was selected for a government pilot programme providing long-term care benefits for the elderly, carried out jointly by the Ministry of Health and Welfare and the National Health Insurance Service. This initiative supports daily living and physical activities for home-based long-term care recipients, with eligible recipients able to purchase or rent devices within an annual budget of 1.6 million won (approximately USD 1,220). This selection represents a significant expansion of institutional support, moving Hyodol from local government welfare distribution into the national long-term care insurance framework.

Public social workers play a critical role in the Hyodol care programme. Research by Shin and Lee (2024) documents how social workers are involved across three programme phases: selecting potential users from among eligible elderly populations, introducing older adults to the robot and facilitating adoption, and maintaining the robotic programme through ongoing monitoring and support. In the Guro District, individual social workers manage up to 200 robots serving 200 seniors, accessing health monitoring data through the companion app. Social workers have reported that the robot's voice analysis and daily check-in data enable earlier detection of welfare concerns, including instances where the system has detected statements indicating suicidal ideation, triggering psychiatric referral. However, social workers also report increased case management workload and emotional labour associated with navigating the care system, though they have maintained high morale regarding the robotic care programme.

Clinical evidence on outcomes comes from a preliminary observational study by Jung et al. (2025) involving 278 community-dwelling older adults (mean age 80.7 years, 88.5% female, 93.1% living alone) in rural and medically underserved areas of South Korea. After approximately six months of Hyodol use, participants showed statistically significant reductions in depressive symptoms as measured by the Geriatric Depression Scale-Short Form (mean score decreased from 6.69 to 5.05, p<0.001), a 45% reduction in the proportion of individuals at high risk of depression, significant improvements in loneliness scores on the UCLA Loneliness Scale (reduced from 44.0 to 41.3 points, p<0.001), and improved medication adherence (increased from 11.2 to 12.2, p<0.001). However, subgroup analysis of 14 clinically diagnosed depressed participants showed no significant improvements, suggesting the robot may be more effectively targeted toward older adults with subclinical or mild depressive symptoms rather than as a clinical intervention for diagnosed depression.

The Hyodol programme operates in the context of South Korea's acute demographic challenge. Approximately 19.2% of the population (9.93 million people) are aged 65 or older, and approximately 10 older adults die by suicide each day. The strained public health system has increasingly turned to technology companies to address what has been termed the 'K-elderly crisis' and fill gaps in the social care workforce. Hyodol won 'Best Mobile Innovation for Connected Health and Wellbeing' at Mobile World Congress Barcelona in February 2024 and was registered with the US FDA in late 2024, with planned US market entry by early 2026.

Classifications follow the DCI AI Hub Taxonomy. Hover over field labels for definitions.

Social Protection Functions

Implementation/delivery chain
Provision of payments/services primaryCase management
SP Pillar (Primary) The social protection branch: social assistance, social insurance, or labour market programmes. Social assistance
Programme Name Hyodol AI Care Robot Distribution Programme for Elderly Living Alone
Programme Type The type of social protection programme, classified under social assistance, social insurance, or labour market programmes. View in glossary Other
System Level Where in the social protection system the AI is applied: policy level, programme design, or implementation/delivery chain. View in glossary Implementation/delivery chain
Programme Description Government-funded distribution of AI companion robots to elderly people living alone through local government welfare centres and, from 2024, the national long-term care insurance system administered by the Ministry of Health and Welfare and National Health Insurance Service. Provides health monitoring, companionship, medication reminders, emergency alerts, and social worker integration for isolated seniors.
Implementation Type How the AI output is produced: Classical ML, Deep learning, Foundation model, or Hybrid. Affects validation, compute requirements, and governance profile. View in glossary Hybrid
Lifecycle Stage Current stage in the AI lifecycle, from problem identification through to monitoring, maintenance and decommissioning. View in glossary Integration and Deployment
Model Provenance Origin of the AI model: developed in-house, adapted from open-source, commercial/proprietary, or accessed via third-party API. View in glossary API-accessed third-party
Compute Environment Where the AI system runs: on-premise, government cloud, commercial cloud, or edge/device. View in glossary Not documented
Compute Provider The specific cloud or infrastructure provider hosting the AI system. OpenAI (ChatGPT API for conversational AI); Microsoft (voice analysis AI). (Rest of World 2025)
Sovereignty Quadrant Classification of data and compute sovereignty: I (Sovereign), II (Federated/Hybrid), III (Cloud with safeguards), or IV (Shared Innovation Zone). View in glossary Not assessed
Hybrid Components ChatGPT-powered conversational AI (foundation model) combined with classical sensor-based monitoring (infrared, touch, microphone) and Microsoft AI voice analysis for mood assessment
Data Residency Where the data used by the AI system is stored: domestic, regional, or international. View in glossary Not documented
Data Residency Detail Additional detail on the specific data hosting arrangements and jurisdictions. CEO states anonymised data stored for three years in cloud; voice recordings used to train chatbot but not sold to third parties. Specific hosting jurisdiction not confirmed. (Rest of World 2025)
Cross-Border Transfer Whether data crosses national borders, and if so, whether documented safeguards are in place. View in glossary Not documented
Is Agentic Whether the system autonomously plans and executes multi-step workflows, selecting tools and chaining actions with limited human intervention. View in glossary Partial
Agentic Pipeline Description of the chained workflow steps in the agentic pipeline. Robot autonomously initiates conversations, delivers medication reminders, conducts daily health check-ins, and triggers emergency alerts based on sensor data (infrared motion detection, fall detection). However, welfare interventions and clinical responses require human social worker action via companion app.
Agentic Autonomy Degree of autonomy: fully autonomous, semi-autonomous (human checkpoints), or supervised (human approval at each step). Semi-autonomous
Override Points Where in the pipeline human review or override is triggered. Social workers receive alerts and monitoring data via smartphone app and decide on welfare interventions; family members receive emergency notifications; psychiatric referral triggered by social worker review of flagged statements, not autonomous robot action.
Decision Criticality The rights impact of the decision the AI supports. High criticality requires HITL oversight; moderate requires HOTL; low may operate HOOTL. View in glossary Moderate
Human Oversight Type Level of human involvement: Human-in-the-Loop (active review), Human-on-the-Loop (monitoring), or Human-out-of-the-Loop (periodic audit). View in glossary HOTL
Development Process Whether the AI system was developed fully in-house, through a mix of in-house and third-party, or fully by an external provider. View in glossary Fully third-party developed
Highest Risk Category The most significant structural risk source identified: data, model, operational, governance, or market/sovereignty risks. View in glossary Market, sovereignty and industry structure risks
Risk Assessment Status Whether a formal risk assessment, informal assessment, or independent audit has been conducted for this system. Not assessed

Risk Dimensions

Data-related risks
Governance and institutional oversight risks
Market, sovereignty and industry structure risks
  • Human oversight protocol
CategorySensitivityCross-System LinkageAvailabilityKey Constraints
Administrative data from other sectorsPersonalSingle source (no linkage)Currently available and usedInfrared motion sensor data, touch sensor data, and fall detection data collected continuously from robot hardware. Transmitted to social workers and family via companion app for remote monitoring.
Unstructured and text-based contentSpecial categorySingle source (no linkage)Currently available and usedVoice recordings of elderly users during daily health check-ins and free conversation; includes mood, pain, medication adherence, and sleep data. CEO states anonymised data stored three years in cloud; voice recordings used to train chatbot. Sensitivity is special category due to health data and potential mental health indicators including suicidal ideation detection.

Aju Press (2024) 'AI robot Hyodol selected for state elderly care project', Aju Press, 30 September. Available at: https://www.ajupress.com/view/20240930154020160 (Accessed: 25 March 2026).

View source News article / media

Jung, H.W., Kim, Y., Kim, H., Kim, M., Lee, H., Park, J.Y., Kim, W.J. and Park, J. (2025) 'Socially Assistive Robot Hyodol for Depressive Symptoms of Community-Dwelling Older Adults in Medically Underserved Areas: A Preliminary Study', Journal of Clinical Medicine, 15(1), p. 217. doi: 10.3390/jcm15010217.

View source Academic journal article

Shin, H. and Lee, O.E. (2024) 'Who is behind the robot? The role of public social workers in implementing robotic eldercare program in South Korea', Social Work in Health Care, 63(4-5), pp. 311-327. doi: 10.1080/00981389.2024.2324849.

View source Academic journal article

Chandran, R. (2025) 'Hyodol AI robots ease loneliness for South Korea's seniors', Rest of World, 14 January. Available at: https://restofworld.org/2025/korea-ai-robot-senior-care-hyodol/ (Accessed: 25 March 2026).

View source News article / media
Deployment Status How far the system has progressed into real-world operational use, from concept/exploration through to scaled and institutionalised. View in glossary Operational Deployment (Limited Rollout)
Year Initiated The year the AI system was first initiated or development began. 2019
Scale / Coverage The scale and geographic or population coverage of the deployment. Over 12,000 units deployed nationally across South Korea as of 2025; over 1,300 units in South Jeolla province; 412 robots distributed in Guro District, Seoul since 2019. (Rest of World 2025)
Funding Source The source(s) of funding for the AI system development and deployment. Local government welfare budgets (e.g. Guro District invested 200 million won / ~USD 143,867 in 2019 with federal industrial technology ministry support); national long-term care insurance budget (1.6 million won / ~USD 1,220 per eligible recipient annually from 2024 pilot). (Rest of World 2025; Aju Press 2024)
Technical Partners External technology vendors, academic partners, or development partners involved. Hyodol Co., Ltd. (robot manufacturer and AI platform developer); OpenAI (ChatGPT API for conversational AI); Microsoft (voice analysis AI programme). (Rest of World 2025)
Outcomes / Results Jung et al. (2025) preliminary study of 278 older adults (mean age 80.7, 93.1% living alone) in medically underserved areas: after ~6 months, GDS-S depression scores decreased from 6.69 to 5.05 (p<0.001); 45% reduction in high-risk depression group; UCLA Loneliness Scale reduced from 44.0 to 41.3 (p<0.001); medication adherence increased from 11.2 to 12.2 (p<0.001). However, subgroup of 14 clinically diagnosed depressed participants showed no significant improvements. User acceptance 2.89/4.0; satisfaction 3.16/4.0. (Jung et al. 2025)
Challenges Increased workload and emotional labour for social workers managing robot programmes (Shin & Lee 2024); concerns about potential deepened isolation through substituting human contact with robot interaction; instances of elderly users taking robot statements literally; single social worker managing up to 200 robots/seniors in Guro District raises capacity concerns. (Rest of World 2025; Shin & Lee 2024)

How to Cite

DCI AI Hub (2026). 'Hyodol AI Care Robot for Elderly Living Alone (South Korea)', AI Hub AI Tracker, case KOR-004. Digital Convergence Initiative. Available at: https://socialprotectionai.org/use-case/KOR-004 [Accessed: 1 April 2026].

Change History

Updated 1 Apr 2026, 08:11
by system (system)
Created 30 Mar 2026, 08:40
by v2-import (import)