The Transformative Potential of Artificial Intelligence in Eldercare: A Comprehensive Analysis

Many thanks to our sponsor Esdebe who helped us prepare this research report.

Abstract

The profound demographic shift towards an aging global population presents an unprecedented set of challenges to existing healthcare and social support infrastructures. Traditional caregiving models are increasingly strained by escalating demand, workforce shortages, and the intricate, multifaceted needs of older adults. In response, Artificial Intelligence (AI) has emerged as a disruptive and transformative force, offering innovative solutions poised to revolutionize the provision of eldercare services. This comprehensive research report delves into the diverse and expanding applications of AI across the eldercare continuum, encompassing areas such as enhancing social connection through AI companions, ensuring safety via advanced fall detection systems, optimizing health management with intelligent medication reminders, fostering independent living through sophisticated smart home environments, and augmenting clinical decision-making with AI-powered diagnostic support tools. The report meticulously examines the fundamental technological pillars underpinning these applications, including machine learning, natural language processing, robotics, computer vision, and the Internet of Things (IoT). Crucially, it undertakes a thorough exploration of the intricate ethical frameworks and complex regulatory challenges inherent in deploying AI within such a sensitive domain, addressing issues of autonomy, privacy, algorithmic bias, and accountability. Finally, it considers the broader societal and economic implications of AI, critically analyzing its potential to redefine established care models, scale services, and ultimately enhance the quality of life, independence, and well-being for older adults globally, while emphasizing the imperative for a human-centered and ethically guided approach.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

1. Introduction: The Imperative for Innovation in Eldercare

The 21st century is characterized by an undeniable global demographic phenomenon: an accelerating shift towards an older population. Projections from organizations such as the United Nations indicate that by 2050, one in six people worldwide will be over age 65, and the number of persons aged 80 years or over is expected to triple, reaching 426 million (United Nations, 2020). This demographic dividend, while a testament to advancements in public health and medicine, simultaneously presents an unprecedented set of challenges to healthcare systems, economies, and social structures across nations. As individuals age, they frequently encounter a natural decline in physical and cognitive abilities, necessitating progressively higher levels of support and specialized care. This includes managing chronic diseases, coping with mobility limitations, addressing cognitive impairments, and combating social isolation (World Health Organization, 2021).

Traditional caregiving models, heavily reliant on human labor and often informal family support, are confronting significant strain. This strain manifests in various forms: severe workforce shortages in professional care sectors, escalating economic burdens associated with long-term care, and the profound emotional and physical demands placed upon informal caregivers (Eldercare Workforce Alliance, n.d.). The complexity of geriatric care further exacerbates these challenges, requiring highly individualized approaches that consider a multitude of factors, from medical conditions and psychological well-being to social determinants of health and personal preferences. The search for sustainable, effective, and humane solutions has become a global imperative.

Within this context, Artificial Intelligence (AI) has rapidly emerged as a transformative technological paradigm with the potential to fundamentally reshape the landscape of eldercare. AI, in its broadest sense, encompasses a diverse array of advanced computational technologies designed to enable machines to perform tasks that typically require human intelligence. This includes learning from data, understanding natural language, recognizing patterns, making predictions, and adapting to new information. Key AI sub-fields pertinent to eldercare include machine learning (ML), natural language processing (NLP), robotics, computer vision, and the Internet of Things (IoT).

These AI capabilities can be harnessed to create innovative solutions that address many of the aforementioned challenges in eldercare. From augmenting human caregivers and providing continuous monitoring to facilitating social engagement and supporting clinical diagnostics, AI applications are remarkably diverse and are evolving at an accelerating pace. However, the integration of AI into such a sensitive and deeply human domain is not without its complexities. It mandates a rigorous examination of critical ethical principles, robust regulatory frameworks, and societal implications to ensure that these technologies are developed and deployed responsibly, equitably, and beneficently. This report aims to provide a comprehensive analysis of these facets, offering a detailed exploration of AI’s current and prospective role in enhancing the quality of life for older adults and redefining the future of caregiving.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

2. Applications of AI in Eldercare: Enhancing Support and Well-being

The applications of Artificial Intelligence in eldercare are multifaceted, extending across prevention, monitoring, assistance, and diagnosis. These technologies are designed to augment human capabilities, promote independence, and improve the overall well-being of older adults.

2.1 AI Companions and Social Engagement

Social isolation and loneliness represent significant public health concerns among older adults, often leading to detrimental health outcomes such as depression, cognitive decline, cardiovascular disease, and increased mortality risk (Holt-Lunstad, 2017). AI companions offer innovative avenues to mitigate these challenges by providing interactive engagement and emotional support.

These AI companions manifest in various forms: conversational AI agents (chatbots), robotic pets, and more sophisticated humanoid robots. At their core, these systems leverage advanced Natural Language Processing (NLP) to facilitate voice-based or text-based interactions, engaging users in meaningful conversations. For example, systems like CareYaya’s QuikTok demonstrate the utility of voice-based AI companions, accessible even via traditional telephone lines, ensuring inclusivity for older adults who may not possess smartphones or internet access (CareYaya Health Technologies, 2024). These companions can discuss a range of topics, provide reminders, play games, and share anecdotes, thereby offering cognitive stimulation and a sense of connection.

Beyond basic conversation, advanced AI companions are being developed to detect subtle changes in user’s speech patterns, tone, or content that might signal cognitive impairment, emotional distress, or other mental health concerns. Utilizing sentiment analysis and machine learning models, these systems can identify markers indicative of depression or anxiety, discreetly alerting caregivers or healthcare professionals if predefined thresholds are met (Artificial Intelligence in Mental Health, 2025). Robotic pets, such as PARO the therapeutic seal, employ AI to mimic animal behaviors, responding to touch and sound, offering tactile comfort and emotional connection without the responsibilities associated with live animals (Shibata, 2012). More advanced social robots, like Pepper, are designed to engage in complex human-robot interactions, capable of recognizing faces, understanding emotions, and adapting their responses to individual user preferences and moods (Pepper Robot, n.d.).

The benefits are substantial: a reduction in reported feelings of loneliness, improved mood, enhanced cognitive engagement, and a continuous, non-judgmental presence. However, the ethical implications surrounding the authenticity of these relationships and the potential for ‘dehumanization’ must be carefully considered, ensuring these companions augment, rather than replace, genuine human interaction.

2.2 Advanced Fall Detection Systems

Falls are a leading cause of injury, disability, and mortality among older adults, often resulting in hip fractures, head trauma, and a significant decline in independent living. The timely detection of falls is critical for minimizing ‘long lie times,’ which are directly correlated with increased morbidity and mortality (Donaldson et al., 2019). AI-powered fall detection systems represent a significant advancement over traditional alert buttons, offering proactive and often automated monitoring.

These systems can be broadly categorized into three types: wearable devices, environmental sensors, and ambient intelligence systems.

• Wearable Devices: These typically include accelerometers, gyroscopes, and magnetometers integrated into smartwatches, pendants, or patches. Machine learning algorithms analyze motion patterns to distinguish between normal activities (e.g., sitting down quickly, bending over) and actual falls, triggering an alert to caregivers or emergency services when a fall is detected. While effective, user adherence can be a challenge, as older adults may forget to wear them or find them uncomfortable.
• Environmental Sensors: These systems are embedded within the living space. They include pressure mats on floors, radar or lidar sensors that detect body movements without requiring line-of-sight (preserving privacy better than cameras), and passive infrared sensors. Camera-based systems, utilizing computer vision algorithms, can analyze posture and movement to detect falls with high accuracy, although they raise significant privacy concerns (Rougier et al., 2011).
• Ambient Intelligence Systems: These integrate various sensor types into a holistic smart home environment, combining data from motion sensors, sound sensors (to detect impact noises or cries for help), and even vital sign monitors. AI algorithms then process this multi-modal data to create a comprehensive picture of the individual’s activity patterns, identifying anomalies that may indicate a fall or other distress, such as prolonged inactivity in an unusual position.

AI’s role is crucial in reducing false positives, a common issue in earlier fall detection technologies, by learning an individual’s unique movement patterns and adapting to their environment. Real-time data processing and predictive analytics can even identify gait instability or increased risk factors for falls, allowing for preventive interventions before an incident occurs (Mubashir et al., 2013). The benefits include rapid response, reduced fear of falling, and enhanced peace of mind for both older adults and their caregivers.

2.3 Intelligent Medication Management and Reminders

Medication non-adherence is a pervasive and costly problem in eldercare, particularly for older adults managing multiple chronic conditions, which often involves complex polypharmacy regimens. Errors in medication intake, such as missed doses, incorrect dosages, or taking medications at the wrong time, can lead to worsening health conditions, adverse drug reactions, and increased hospitalizations. AI-driven medication reminder systems offer sophisticated solutions to improve adherence and enhance medication safety.

These systems go beyond simple alarms. They integrate various features:

• Personalized Alerts: Utilizing machine learning, systems can learn a user’s daily routine and preferences to deliver reminders via multiple modalities (visual displays, auditory alerts, haptic feedback on wearables, voice assistants) at optimal times, adapting to an individual’s schedule rather than a rigid timetable.
• Smart Dispensers: AI-integrated smart pill dispensers automatically sort and dispense medications at prescribed times, often requiring a physical action (e.g., pressing a button) to confirm intake. Some advanced systems can even detect if medication has been removed but not consumed.
• Tracking and Reporting: The systems log medication intake, providing a comprehensive adherence record for both the user and remote caregivers or healthcare providers. This data can be invaluable for clinicians to assess treatment effectiveness and adjust prescriptions.
• Drug Interaction and Side Effect Monitoring: Sophisticated AI algorithms can cross-reference prescribed medications with a comprehensive drug database, alerting users and caregivers to potential adverse drug interactions or common side effects, and advising on appropriate actions (e.g., ‘take with food’ or ‘avoid driving’).
• Educational Support: AI can provide on-demand, personalized educational information about each medication, its purpose, dosage, and potential effects, improving health literacy and empowering older adults in their self-management (Lee et al., 2018).

Future advancements may include AI-driven dosage adjustments (under strict medical supervision) based on real-time physiological data and personalized adherence strategies informed by behavioral science. The primary benefits are improved medication adherence, reduced adverse events, better management of chronic conditions, and a significant reduction in caregiver burden associated with medication oversight.

2.4 Smart Home Systems for Ambient Assisted Living (AAL)

Smart home technologies, deeply integrated with AI, are central to the concept of Ambient Assisted Living (AAL), which aims to enhance the quality of life for older adults by enabling them to live independently and safely in their own homes for longer. These systems automate daily tasks, continuously monitor environmental conditions, and provide a comprehensive safety net.

Key components of AI-powered smart home systems include:

• Environmental Control: AI algorithms learn occupant preferences and routines to automate lighting, heating, ventilation, and air conditioning (HVAC) systems. For instance, lights can turn on automatically as an individual enters a room and adjust brightness based on ambient light levels and time of day. Thermostats can adapt to comfort preferences and optimize energy consumption.
• Security and Safety Monitoring: AI can detect anomalies such as open doors or windows, unusual noises, gas leaks, or smoke, triggering immediate alerts. Integrated security cameras (with privacy-preserving AI analysis) can identify intruders or unusual activity patterns (e.g., an individual leaving the house at an uncharacteristic hour and not returning).
• Activity Monitoring: Beyond fall detection, AI can track daily activity patterns using a network of non-intrusive sensors (motion, door/window contacts, bed sensors). Deviations from established routines – such as prolonged inactivity, changes in sleep patterns, or unusual bathroom visits – can trigger alerts to caregivers, signaling potential health issues or distress (Rashidi & Mihailidis, 2013).
• Voice Assistants and Smart Appliances: Voice-activated AI assistants (e.g., Amazon Alexa, Google Assistant) integrate with smart home devices, allowing older adults to control their environment, make calls, access information, or even receive cognitive prompts using natural language. Smart appliances can simplify tasks, such as smart ovens with automatic shut-off features or smart refrigerators that monitor food inventory and suggest meal plans.
• Integration with Wearable Devices: Smart home systems often seamlessly integrate with wearable health trackers, aggregating data on vital signs (heart rate, blood pressure, sleep quality), activity levels, and glucose readings. AI then analyzes this holistic dataset to provide a comprehensive view of an individual’s health status and alert to any significant deviations.

These integrated systems foster greater independence, enhance personal safety, reduce caregiver burden by providing remote oversight, and can significantly improve the older adult’s overall quality of life by making their living environment more responsive and supportive.

2.5 AI-Assisted Diagnostic and Clinical Support

AI applications in diagnostic support represent a paradigm shift in how healthcare professionals identify, diagnose, and manage health conditions pertinent to older adults. Machine learning algorithms, particularly deep learning, excel at pattern recognition in vast and complex datasets, making them invaluable tools in clinical decision-making.

• Early Detection of Neurodegenerative Diseases: AI is proving instrumental in the early detection of conditions like Alzheimer’s Disease and Parkinson’s Disease. For Alzheimer’s, AI can analyze subtle changes in brain MRI or PET scans, cognitive test scores, speech patterns (e.g., word-finding difficulties, changes in coherence), and even gait analysis to identify biomarkers indicative of early-stage disease long before clinical symptoms become pronounced (Liu et al., 2019). For Parkinson’s, AI can analyze tremors, speech dysarthria, and even handwriting samples to aid in early diagnosis and monitor disease progression.
• Cancer Diagnosis and Prognosis: In oncology, AI algorithms can analyze medical imaging (mammograms, CT scans, MRIs, pathological slides) with accuracy comparable to, or even exceeding, human experts, identifying subtle abnormalities that might be missed by the human eye. This applies to various cancers prevalent in older populations, such as breast, lung, colorectal, and prostate cancer. AI can also analyze genomic data and electronic health records to predict treatment response and disease recurrence, personalizing prognosis and therapeutic strategies (Topol, 2019).
• Cardiovascular Disease Risk Prediction: AI can analyze vast amounts of patient data – including ECG readings, blood test results, lifestyle factors, and genetic predispositions – to predict the risk of heart attacks, strokes, and other cardiovascular events, allowing for proactive preventive interventions.
• Electronic Health Record (EHR) Analysis: NLP-powered AI can extract critical information from unstructured clinical notes, aiding in the identification of complex patient cohorts, flagging potential drug interactions, and ensuring adherence to clinical guidelines. This reduces the burden on clinicians and enhances the quality of data for research and care planning.
• Decision Support Systems: AI-driven clinical decision support systems (CDSS) provide clinicians with evidence-based recommendations for diagnosis, treatment options, and care pathways, considering the individual patient’s comorbidities and unique circumstances. These systems act as intelligent assistants, reducing diagnostic errors and supporting personalized care planning (Artificial Intelligence-Based Clinical Decision Support Systems in Geriatrics, 2023).

By augmenting the capabilities of healthcare professionals, AI diagnostic tools can improve diagnostic accuracy, expedite the diagnostic process, reduce healthcare costs associated with delayed diagnosis, and ultimately lead to more effective and personalized care for older adults.

2.6 Personalized Rehabilitation and Therapy

AI also offers significant potential in the realm of rehabilitation and physical therapy for older adults. After injuries (e.g., falls) or medical events (e.g., stroke), personalized and consistent therapy is crucial for recovery. AI can facilitate this by:

• AI-Guided Exercises: Systems using computer vision (via cameras) or wearable sensors can monitor an individual’s movements during exercises, providing real-time feedback on form and technique, ensuring exercises are performed correctly and safely. This is particularly valuable for home-based rehabilitation, where direct supervision may be limited.
• Virtual Reality (VR) Therapy: AI-powered VR environments can create engaging and immersive therapy sessions, making rehabilitation more enjoyable and motivating. These systems can track progress, adjust difficulty levels dynamically, and provide objective performance metrics, aiding therapists in tailoring treatment plans (Rizzo & Koenig, 2017).
• Adaptive Robotics for Mobility: Exoskeletons and robotic assistance devices, powered by AI, can aid individuals with mobility impairments, helping them to walk or perform daily tasks, while adaptively learning and responding to their specific needs and movements. This can significantly improve functional independence and quality of life.

2.7 Nutritional Monitoring and Meal Planning

Proper nutrition is fundamental to maintaining health and preventing disease in older adults. AI can assist in this area by:

• Dietary Assessment: AI-powered applications, often integrated with smart kitchen appliances or photo recognition (computer vision), can help individuals track their food intake, analyze nutritional content, and identify deficiencies or excesses.
• Personalized Meal Recommendations: Based on an individual’s health conditions (e.g., diabetes, hypertension), allergies, dietary preferences, and activity levels, AI can generate personalized meal plans and recipes. These systems can also remind users about hydration and suggest appropriate portion sizes.
• Grocery List Generation: By linking with smart refrigerators or pantry inventories, AI can automatically generate grocery lists, ensuring older adults have access to necessary ingredients for their recommended diets.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

3. Underlying Technologies: The Foundation of AI in Eldercare

The efficacy and sophistication of AI applications in eldercare are intrinsically linked to advancements in several core technological domains. These underlying technologies provide the computational power, analytical capabilities, and interactive functionalities necessary for intelligent systems.

3.1 Machine Learning (ML)

Machine learning is a cornerstone of modern AI, enabling systems to learn from data, identify patterns, and make predictions or decisions without being explicitly programmed for every specific task. Its application in eldercare is pervasive:

• Supervised Learning: This involves training models on labeled datasets (e.g., images labeled ‘fall’ or ‘no fall,’ medical records with diagnosed conditions). Algorithms like Support Vector Machines (SVMs), Random Forests, and Artificial Neural Networks (ANNs) are used for classification (e.g., diagnosing a disease) or regression (e.g., predicting disease progression or fall risk scores). For instance, an ML model can be trained on a dataset of vital signs and clinical outcomes to predict the likelihood of hospital readmission for an older patient with heart failure.
• Unsupervised Learning: This approach uncovers hidden patterns or structures within unlabeled data. Clustering algorithms can group older adults with similar health profiles or behavioral patterns, allowing for more targeted care interventions or personalized wellness programs. Anomaly detection, a form of unsupervised learning, is crucial in smart home systems for identifying unusual activity patterns that might indicate a problem.
• Reinforcement Learning (RL): RL algorithms learn by interacting with an environment, receiving rewards or penalties for their actions. While less common in current mainstream eldercare applications, RL holds promise for training adaptive robots to perform complex tasks or for personalizing therapeutic interventions based on real-time user responses (Sutton & Barto, 2018).
• Deep Learning: A subfield of ML, deep learning utilizes multi-layered neural networks (deep neural networks) to learn intricate patterns from vast amounts of data, particularly effective for complex data types like images, video, and audio. Convolutional Neural Networks (CNNs) are extensively used for image analysis in diagnostic support (e.g., detecting tumors in medical scans or identifying fall postures), while Recurrent Neural Networks (RNNs) and their variants (LSTMs, GRUs) are powerful for analyzing sequential data such as speech patterns or time-series sensor data in fall detection or activity monitoring.

ML’s strength lies in its ability to continuously improve its performance as more data becomes available, making it ideal for dynamic environments like healthcare where individual needs and conditions evolve.

3.2 Natural Language Processing (NLP)

Natural Language Processing (NLP) empowers machines to understand, interpret, and generate human language in a valuable and meaningful way. This capability is vital for creating intuitive and accessible AI systems for older adults.

• Speech Recognition (Automatic Speech Recognition – ASR): This converts spoken language into text, enabling voice-controlled interfaces for AI companions, smart home systems, and medication reminders. Advanced ASR systems are trained to recognize various accents, speech impediments, and background noise, which is particularly important for older users.
• Natural Language Understanding (NLU): NLU allows AI systems to comprehend the meaning, intent, and context of human language. This is crucial for AI companions to engage in relevant conversations, interpret user requests accurately, and detect emotional states through linguistic cues. For instance, an AI companion can distinguish between ‘I’m feeling down’ and ‘I fell down’ and respond appropriately.
• Natural Language Generation (NLG): NLG enables AI systems to produce human-like text or speech, allowing AI companions to generate coherent and contextually appropriate responses, or for medication reminder systems to verbally explain drug information.
• Sentiment Analysis: A specialized application of NLP, sentiment analysis identifies and extracts subjective information (emotions, opinions, attitudes) from text or speech. In eldercare, it helps AI companions detect signs of loneliness, sadness, or frustration, enabling proactive interventions by alerting caregivers or suggesting mood-lifting activities (Cambria et al., 2017).
• Clinical Text Mining: NLP is used to process vast amounts of unstructured text in Electronic Health Records (EHRs), such as doctor’s notes, discharge summaries, and radiology reports, to extract critical clinical information, identify disease phenotypes, or flag potential care gaps, thereby enhancing diagnostic support and research capabilities.

3.3 Robotics

Robotics, when integrated with AI, provides a physical dimension to eldercare assistance, offering tangible support and companionship. The field encompasses a wide array of robotic types, each with specific applications:

• Social Robots: Designed primarily for emotional and social interaction. Examples include PARO (a therapeutic robotic seal) providing tactile comfort and reducing stress, and humanoid robots like Pepper or Nao, which can engage in conversation, play games, and provide reminders. These robots leverage NLP for communication and computer vision for facial recognition and emotion detection, adapting their behavior to the user (Shibata, 2012).
• Assistive Robots: These robots provide physical aid. This includes mobility assistance robots (e.g., robotic walkers or wheelchairs that navigate autonomously), lifting aid robots to assist caregivers with transfers, and robotic manipulators that can fetch objects, open doors, or perform light household chores. AI enables these robots to perceive their environment, navigate safely, and adapt their movements to user commands and changing situations.
• Service Robots: These are designed for more mundane, repetitive tasks, freeing up human caregivers for more complex emotional or clinical care. Examples include robotic vacuum cleaners, automatic floor scrubbers, or even advanced kitchen robots capable of preparing simple meals.
• Telepresence Robots: These robots allow remote family members or healthcare professionals to ‘visit’ and interact with older adults in their homes, bridging geographical distances and fostering connection through a mobile, interactive interface.

Key robotic capabilities include sophisticated locomotion, object recognition and manipulation, safe Human-Robot Interaction (HRI) design, and continuous learning from user behaviors and environmental feedback. Challenges remain in terms of cost, safety assurances, and ensuring public acceptance and integration into daily life.

3.4 Computer Vision

Computer vision is an AI field that enables computers to ‘see’ and interpret visual information from the world, much like human sight. Its applications in eldercare are critical for monitoring, safety, and interaction:

• Activity Recognition and Monitoring: Using cameras (with privacy safeguards) or depth sensors, computer vision algorithms can track an older adult’s activities within their home. This includes recognizing daily routines, identifying prolonged inactivity, or detecting unusual behaviors that might indicate a problem. For instance, a system can distinguish between a user getting out of bed and going to the bathroom versus lying on the floor.
• Fall Detection: As previously discussed, camera-based fall detection systems employ computer vision to analyze posture, movement trajectories, and velocity to accurately identify falls and distinguish them from non-fall events (Rougier et al., 2011).
• Gait Analysis: Computer vision can be used to analyze an individual’s gait parameters (stride length, speed, balance) over time, identifying subtle changes that could indicate an increased risk of falling or the progression of neurological conditions like Parkinson’s disease.
• Facial Expression and Emotion Recognition: While ethically sensitive, computer vision can analyze facial expressions to infer emotional states, which can be useful for social robots or AI companions to better understand and respond to an older adult’s mood, potentially detecting signs of distress or loneliness.
• Object Recognition: In assistive robotics or smart homes, computer vision helps robots identify objects (e.g., fetching a specific item) or ensures that walking paths are clear of obstacles.

3.5 Internet of Things (IoT)

The Internet of Things (IoT) provides the essential infrastructure for collecting the vast amounts of real-time data that AI systems require for effective operation in eldercare. IoT refers to the network of physical objects embedded with sensors, software, and other technologies for the purpose of connecting and exchanging data with other devices and systems over the internet.

• Sensor Networks: IoT devices form dense sensor networks in eldercare environments. These include wearable health trackers (measuring heart rate, blood pressure, sleep, activity), environmental sensors (temperature, humidity, air quality, gas leaks), motion sensors, bed sensors, door/window sensors, and smart appliances.
• Data Collection and Transmission: IoT devices continuously collect data, which is then transmitted wirelessly (via Wi-Fi, Bluetooth, Zigbee, cellular networks) to cloud-based platforms for storage and processing. This constant stream of data forms the raw material for AI algorithms to analyze and derive insights.
• Interoperability: A key challenge and focus area for IoT in eldercare is ensuring interoperability – the ability of different devices and systems from various manufacturers to communicate and exchange data seamlessly. Standardized communication protocols are essential for building integrated eldercare solutions.
• Edge Computing: With the proliferation of IoT devices, processing data closer to the source (at the ‘edge’ of the network) rather than sending all data to the cloud becomes crucial for real-time applications like fall detection. Edge computing reduces latency and bandwidth usage, while also enhancing privacy by processing sensitive data locally (AI and Ethical Robotics Set to Transform Elder Care Services, 2023).

Together, these underlying technologies form a powerful ecosystem, enabling AI to gather, process, and act upon information in complex and intelligent ways, transforming the possibilities within eldercare.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

4. Ethical Considerations: Navigating the Moral Landscape of AI in Eldercare

The integration of Artificial Intelligence into the deeply personal and vulnerable domain of eldercare raises profound ethical questions that demand meticulous consideration. While promising immense benefits, AI technologies also present potential risks if not designed and implemented with a human-centered, ethically guided approach (Ethics of Artificial Intelligence, 2025; MEEGLE, n.d.).

4.1 Autonomy and Informed Consent

Ensuring the autonomy of older adults, particularly those with declining cognitive abilities, is paramount. AI systems should be designed to support, not undermine, an individual’s capacity for self-determination and decision-making. The concept of informed consent becomes particularly complex when dealing with vulnerable populations who may have varying degrees of cognitive impairment.

• Capacity for Consent: How is consent obtained when an older adult’s cognitive function fluctuates or is permanently diminished? This requires careful consideration of legal and ethical frameworks for proxy decision-making or ‘digital guardianship,’ ensuring that a designated representative (family member, legal guardian) makes decisions in the individual’s best interest, aligning with their known values and preferences.
• Transparency and Understanding: Older adults and their caregivers must be fully informed about how AI systems operate, what data they collect, how decisions are made, and what the benefits and risks are. The technical complexity of AI necessitates clear, understandable explanations, avoiding jargon, to ensure truly informed consent. A lack of transparency can erode trust and diminish an individual’s sense of control over their own care.
• User Control and Customization: AI systems should offer customizable settings, allowing users to control the level of monitoring, interaction, and data sharing. The ability to ‘opt-out’ of certain features or disconnect from the system should always be available, reinforcing autonomy.

4.2 Privacy and Data Security

AI applications in eldercare inherently involve the collection, processing, and storage of highly sensitive personal data, including detailed health information, behavioral patterns, location data, and even emotional states. Safeguarding this data against unauthorized access, misuse, and breaches is a critical ethical and legal imperative.

• Types of Data: The breadth of data collected by AI in eldercare is extensive – from vital signs and medication adherence to sleep patterns, conversational transcripts, and even inferred emotional states. This ‘digital footprint’ provides an incredibly intimate picture of an individual’s life.
• Regulatory Compliance: Adherence to robust data protection regulations, such as the Health Insurance Portability and Accountability Act (HIPAA) in the United States and the General Data Protection Regulation (GDPR) in the European Union, is essential. These regulations dictate strict rules for data collection, storage, processing, and sharing, requiring explicit consent and robust security measures.
• Cybersecurity Risks: AI systems are targets for cyberattacks. Data breaches can expose sensitive health information, leading to identity theft, discrimination, or exploitation. Robust encryption, secure data storage, access controls, and regular security audits are vital.
• Data Usage and Sharing: Clear and explicit policies must govern how collected data is used (e.g., for direct care, research, product improvement) and with whom it is shared (e.g., family caregivers, healthcare providers, third-party developers). Data anonymization and de-identification techniques are crucial when data is used for research or system improvement to protect individual identities (Legal and Ethical Considerations of Artificial Intelligence for Residents, 2023).

4.3 Algorithmic Bias and Equity

AI systems learn from the data they are trained on. If this training data is unrepresentative, incomplete, or reflects existing societal biases, the AI system can inadvertently perpetuate or even amplify these biases, leading to unfair, discriminatory, or suboptimal outcomes for certain groups of older adults (Navigating Ethical Considerations in the Use of Artificial Intelligence, 2023).

• Sources of Bias: Bias can stem from various sources: demographic underrepresentation in training datasets (e.g., AI models trained primarily on data from younger, healthier, or specific ethnic groups may perform poorly for older adults or minority populations), historical inequities reflected in medical records, or flawed data collection methodologies.
• Consequences of Bias: In eldercare, algorithmic bias could manifest as: misdiagnosis or delayed diagnosis for certain demographics, inequitable allocation of care resources, ineffective personalized care plans, or even a lack of accessibility for certain linguistic or cultural groups. For instance, a fall detection system trained predominantly on data from Caucasian individuals might perform less accurately for individuals of different body types or skin tones.
• Mitigation Strategies: Addressing algorithmic bias requires proactive measures: developing and utilizing diverse and representative datasets, employing fairness-aware machine learning techniques, continuous monitoring and auditing of AI system performance across different demographic groups, and investing in Explainable AI (XAI) to understand why an AI makes a particular decision. The involvement of multidisciplinary teams, including ethicists, social scientists, and older adult representatives, in the development process is crucial.

4.4 Dependency and Dehumanization

While AI aims to augment care, there is a legitimate concern that an over-reliance on technological solutions could reduce essential human interaction, potentially exacerbating feelings of isolation and leading to ‘dehumanization’ of care.

• Erosion of Human Connection: The unique emotional support, empathy, and intuitive understanding provided by human caregivers are difficult, if not impossible, for current AI systems to replicate. Substituting human touch with robotic interaction, especially in areas of intimate care or emotional support, could lead to a perceived loss of dignity and genuine connection.
• Psychological Impact: Older adults may feel constantly monitored or judged by AI systems, leading to a sense of diminished privacy or control. There is also a risk of passive engagement, where individuals become overly reliant on AI for decision-making or daily tasks, potentially reducing their own cognitive and physical activity.
• Maintaining Human Touch: It is imperative that AI in eldercare is viewed as a complement to human caregivers, not a replacement. AI can handle routine, repetitive, or data-intensive tasks, thereby freeing human caregivers to focus on complex clinical needs, emotional support, and fostering meaningful relationships. Hybrid care models that integrate technology with robust human interaction are the ethical imperative.

4.5 Equity and Accessibility

Beyond algorithmic bias, there are broader ethical considerations regarding equitable access to AI technologies in eldercare.

• Digital Divide: Not all older adults have access to the internet, smartphones, or the financial resources to acquire advanced AI-powered devices. Socioeconomic disparities and geographical location can create a ‘digital divide,’ preventing vulnerable populations from benefiting from these innovations.
• Technological Literacy: Many older adults may lack the technological literacy or confidence to effectively interact with complex AI systems. User interfaces must be designed with extreme simplicity, intuitiveness, and accessibility in mind.
• Cost: High development and implementation costs can make advanced AI eldercare solutions expensive, potentially limiting their availability to wealthier individuals or communities. Policies and funding models are needed to ensure equitable access across all socioeconomic strata.

4.6 Trust and Explainable AI (XAI)

For AI systems to be truly beneficial and accepted, older adults, their families, and caregivers must trust them. This trust is contingent on understanding how these systems arrive at their decisions and predictions.

• The ‘Black Box’ Problem: Many advanced AI algorithms, particularly deep learning models, operate as ‘black boxes,’ where their internal decision-making processes are opaque and difficult for humans to interpret. This lack of transparency can lead to distrust, especially when critical health decisions are involved.
• Explainable AI (XAI): XAI aims to make AI models more transparent and interpretable, allowing users to understand the rationale behind an AI’s output. In eldercare, XAI can provide explanations for a diagnostic recommendation, a fall risk assessment, or a behavioral alert, building confidence and facilitating informed human oversight.

Addressing these ethical considerations requires a concerted effort from AI developers, policymakers, healthcare providers, ethicists, and older adults themselves. Co-design and participatory approaches are crucial to ensure that AI technologies truly serve the best interests of the aging population.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

5. Regulatory Challenges: Establishing a Framework for Responsible Innovation

The rapid pace of AI development and its deployment into the highly regulated healthcare sector, particularly for vulnerable populations, presents a complex web of regulatory challenges. Existing legal frameworks often struggle to keep pace with technological advancements, necessitating new approaches to ensure safety, efficacy, accountability, and ethical adherence (Legal and Ethical Considerations of Artificial Intelligence for Residents, 2023).

5.1 Standardization and Interoperability

Establishing standardized protocols for the development, testing, and implementation of AI systems in eldercare is critical for ensuring consistency, reliability, and safety across different products and providers.

• Technical Standards: There is a pressing need for standards governing AI system performance, accuracy, robustness, and cybersecurity. This includes benchmarks for fall detection accuracy, reliability of medication reminders, and safety protocols for human-robot interaction. Organizations like the International Organization for Standardization (ISO) and the Institute of Electrical and Electronics Engineers (IEEE) are actively developing AI ethics and safety standards.
• Interoperability: For AI systems to truly create integrated care ecosystems, different devices, platforms, and data sources (e.g., smart home sensors, wearables, EHRs) must be able to communicate and exchange data seamlessly. Standardized data formats and communication protocols, such as Fast Healthcare Interoperability Resources (FHIR), are essential to prevent data silos and enable holistic AI analysis.
• Ethical Guidelines as Standards: Beyond technical standards, ethical principles (e.g., fairness, transparency, privacy) need to be translated into actionable guidelines and ‘standards of care’ for AI design, deployment, and oversight. This includes requirements for bias detection and mitigation, clear consent mechanisms, and user control features.

5.2 Accountability and Liability

Determining accountability when AI systems are involved in healthcare decisions or incidents is one of the most complex regulatory challenges. Traditional legal frameworks, designed for human actions and responsibilities, struggle to assign liability in the context of autonomous AI.

• Defining Responsibility: Who is accountable if an AI diagnostic tool makes an error leading to patient harm? Is it the AI developer, the healthcare provider who used the tool, the hospital, or a combination? Legal frameworks must evolve to delineate the responsibilities of AI manufacturers, software developers, healthcare institutions, clinicians using AI tools, and even the end-user or caregiver.
• Medical Malpractice: The concept of medical malpractice needs to be re-evaluated for AI-assisted care. If a clinician follows an AI recommendation that proves to be incorrect, does liability shift? This requires clear guidelines on the expected level of human oversight and critical evaluation of AI outputs.
• Product Liability: If a robotic companion causes an injury due to a software glitch, does product liability law apply, and to what extent? The complexity of AI’s adaptive learning capabilities further complicates this, as the system’s behavior may evolve beyond its initial programming.
• Insurance Implications: The introduction of AI will necessitate new insurance models to cover the risks associated with AI failures or errors in eldercare. This includes liability insurance for AI developers and healthcare providers utilizing AI.

5.3 Continuous Monitoring and Evaluation

Unlike traditional medical devices that undergo a static approval process, AI systems, particularly those using machine learning, are often adaptive and can change their behavior over time as they learn from new data or receive updates. This necessitates a framework for continuous monitoring and re-evaluation.

• Post-Market Surveillance: Regulatory bodies (e.g., FDA in the US, EMA in Europe) need robust post-market surveillance systems specifically tailored for AI, capable of tracking the performance, safety, and ethical compliance of AI systems in real-world settings over their operational lifespan.
• Version Control and Updates: AI models are frequently updated. Regulations must address how these updates are managed, tested, and approved, especially if they alter the core functionality or decision-making logic of the system, to ensure continued safety and efficacy.
• Data Governance: A continuous data governance framework is required to ensure that the data used for ongoing AI training and operation remains high-quality, unbiased, and compliant with privacy regulations. This includes processes for data collection, storage, anonymization, and auditing.
• Ethical Oversight Committees: Ongoing ethical review by multidisciplinary committees, involving ethicists, clinicians, technologists, and patient advocates, can provide a crucial layer of oversight for AI systems, particularly as they adapt and learn.

5.4 Certification and Approval Processes

The regulatory pathway for AI in healthcare is still nascent and often relies on adapting existing frameworks for medical devices or software. However, the unique characteristics of AI necessitate specialized approaches.

• Adaptive AI: Regulators face the challenge of approving AI systems that can change post-deployment. This may require ‘locked’ algorithms for certain critical functions or stringent requirements for real-world evidence collection and algorithmic transparency for adaptive systems.
• Software as a Medical Device (SaMD): Many AI applications fall under the SaMD framework, but the specific requirements for validation, verification, and clinical evidence generation for AI algorithms are still being defined by regulatory bodies.
• Clinical Validation: Rigorous clinical trials and real-world evidence studies are essential to demonstrate the safety and efficacy of AI in eldercare, ensuring that these tools provide tangible benefits and do not introduce unintended harms.

5.5 Public Engagement and Education

An often-overlooked regulatory challenge is the need to foster public understanding and trust in AI. Regulations should encourage transparency and education.

• Informed Public: Regulatory efforts should extend to educating older adults, their families, and caregivers about the capabilities, limitations, and ethical considerations of AI. This empowers individuals to make informed decisions about adopting or interacting with these technologies.
• Stakeholder Involvement: Regulations should mandate active involvement of end-users, patient advocacy groups, and care organizations in the design and evaluation process of AI tools, ensuring that solutions are truly user-centered and meet real-world needs.

Developing a robust and agile regulatory framework for AI in eldercare is a continuous process that demands collaboration among governments, industry, academia, healthcare providers, and civil society. It is a critical step towards realizing the benefits of AI while mitigating its inherent risks.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

6. Broader Implications of AI in Eldercare: Reshaping the Future of Care

The integration of Artificial Intelligence into eldercare extends far beyond individual applications, holding profound broader implications for healthcare systems, economies, and societies at large. It heralds a significant shift in how care is conceptualized, delivered, and experienced (Applications of Artificial Intelligence, 2025; TMA Solutions, n.d.).

6.1 Scalability and Addressing Workforce Shortages

One of the most compelling arguments for AI in eldercare lies in its inherent scalability. The global aging demographic is creating a growing ‘care gap’ – a widening disparity between the demand for care services and the available human workforce (Eldercare Workforce Alliance, n.d.). AI offers a potent solution to this looming crisis.

• Augmenting Human Caregivers: AI technologies can automate or assist with routine, repetitive, and data-intensive tasks, thereby freeing human caregivers to focus on more complex clinical needs, emotional support, and personalized human interaction. For example, AI-powered scheduling, medication reminders, and vital sign monitoring can reduce the administrative burden on nurses and personal care aides.
• Expanding Reach: AI-driven remote patient monitoring, tele-health platforms, and AI companions can extend care services to a larger number of individuals, including those in remote or underserved areas, without requiring a proportional increase in human personnel. This democratizes access to certain levels of care and support.
• Economic Efficiency: By optimizing resource allocation, reducing hospital readmissions through proactive monitoring, and enabling older adults to age in place for longer, AI can contribute to significant cost savings within healthcare systems, making care more economically sustainable in the long run.

6.2 Proactive and Predictive Health Monitoring

AI fundamentally shifts the paradigm of eldercare from reactive to proactive and preventive. Instead of responding to crises, AI enables continuous, real-time health monitoring and predictive analytics, facilitating early detection and timely intervention.

• Early Warning Systems: AI-powered sensors in smart homes and wearables continuously collect data on vital signs, activity patterns, sleep quality, and even subtle changes in gait or speech. Machine learning algorithms can analyze this vast data stream to detect anomalies or predict potential health deteriorations (e.g., an impending fall, the onset of an infection, or a exacerbation of a chronic condition) before they become critical.
• Reduced Hospitalizations: Proactive monitoring and early interventions can significantly reduce the incidence of preventable hospitalizations and emergency room visits, which are costly and often disruptive for older adults.
• Improved Quality of Life: By mitigating health risks before they escalate, AI contributes to better health outcomes, greater functional independence, and an enhanced quality of life for older adults, allowing them to remain active and engaged for longer.

6.3 Personalized and Adaptive Information and Care Delivery

AI excels at personalization, tailoring information, services, and interventions to the unique preferences, needs, and evolving health status of each individual older adult. This moves away from a ‘one-size-fits-all’ approach to highly individualized care.

• Adaptive Learning: AI systems can continuously learn from an individual’s interactions, behaviors, and health data to refine their services. An AI companion can adapt its conversational style based on user preferences, a smart home system can optimize environmental controls based on learned routines, and a rehabilitation robot can adjust exercise difficulty based on real-time performance.
• Targeted Information: AI can deliver highly relevant health information, medication reminders, or activity suggestions that are specific to an individual’s conditions, cognitive abilities, and interests, enhancing engagement and adherence.
• Empowerment and Self-Management: By providing accessible, personalized information and support, AI empowers older adults to take a more active role in managing their own health and well-being, fostering a sense of control and independence.

6.4 Redefining Care Models and the Role of Caregivers

AI’s integration challenges traditional caregiving models, prompting a necessary re-evaluation of roles, responsibilities, and organizational structures within eldercare.

• Hybrid Care Models: The future of eldercare will likely involve ‘hybrid care models’ that seamlessly integrate human expertise with technological support. This means human caregivers focus on complex clinical tasks, emotional connection, and holistic well-being, while AI handles monitoring, reminders, and routine assistance. This can lead to a more efficient, effective, and humane care system.
• Evolution of Caregiver Roles: AI will transform the role of human caregivers from primarily performing physical tasks to becoming supervisors of AI systems, data interpreters, and facilitators of technology-enhanced care. This requires new skills, training, and a collaborative mindset.
• Aging in Place: AI significantly strengthens the feasibility of ‘aging in place’ – allowing older adults to remain in their homes and communities for as long as possible. Smart homes, remote monitoring, and AI companions provide the infrastructure for sustained independent living, reducing the need for institutional care.
• Integrated Ecosystems of Care: AI can act as the connective tissue between various components of the care ecosystem – home, primary care, specialists, family caregivers, and community services – creating a more coordinated and holistic approach to eldercare.

6.5 Economic and Societal Impact

Beyond care delivery, AI in eldercare will have broader economic and societal repercussions.

• Job Creation and Evolution: While some fear job displacement, AI is more likely to create new roles (e.g., AI system technicians, data analysts for eldercare, AI ethics specialists) and elevate existing caregiving roles, requiring new skill sets and potentially better wages.
• Investment and Innovation: The demand for AI in eldercare will stimulate significant investment in research and development, fostering innovation in gerontechnology and geriatric medicine.
• Shifting Perceptions of Aging: By enabling greater independence and quality of life, AI can contribute to a more positive societal perception of aging, viewing older adults as active, contributing members of society rather than merely recipients of care.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

7. Conclusion: Charting a Human-Centered Future for AI in Eldercare

Artificial Intelligence stands at the precipice of profoundly transforming eldercare, offering unprecedented opportunities to enhance the quality, efficiency, and personalization of services for a rapidly expanding global aging population. From mitigating the profound impact of social isolation through empathetic AI companions and safeguarding against falls with advanced detection systems, to optimizing medication adherence, fostering independent living via intelligent smart homes, and revolutionizing diagnostic support for critical conditions, AI’s potential to significantly improve the lives of older adults is undeniable. The foundational technological advancements in machine learning, natural language processing, robotics, computer vision, and the Internet of Things are continuously expanding the horizon of what is achievable, promising a future where technology acts as a powerful enabler for well-being and autonomy.

However, the integration of AI into this deeply sensitive and personal domain is not merely a technological challenge; it is fundamentally an ethical and societal undertaking. This report has underscored the imperative of approaching AI deployment with rigorous ethical consideration, addressing critical issues such as safeguarding autonomy and ensuring informed consent, protecting the privacy and security of sensitive personal data, actively mitigating algorithmic biases to ensure equitable care, and carefully managing the potential for over-dependency or dehumanization of care. Concurrently, the nascent and evolving regulatory landscape demands robust frameworks for standardization, accountability, continuous monitoring, and effective certification processes to ensure the responsible, safe, and trustworthy development and deployment of these powerful tools.

Ultimately, the promise of AI in eldercare is not in replacing the invaluable human element of compassion and connection, but in augmenting it. AI’s role is to free human caregivers from routine and arduous tasks, allowing them to dedicate more time to the complex emotional, social, and nuanced clinical needs of older adults. It is about creating hybrid care models that leverage the strengths of both human empathy and technological efficiency, fostering environments where older adults can age with dignity, independence, and a sustained quality of life.

Realizing this transformative potential hinges on a collaborative and human-centered approach. This necessitates ongoing dialogue and partnership among AI developers, healthcare providers, policymakers, ethicists, researchers, and, crucially, older adults themselves and their families. By prioritizing ethical design, ensuring robust regulatory oversight, promoting accessibility, and fostering genuine trust, Artificial Intelligence can indeed play a pivotal and benevolent role in shaping a future where aging is synonymous with empowered living and comprehensive, dignified care.

Many thanks to our sponsor Esdebe who helped us prepare this research report.

References

• Artificial Intelligence and Assistive Robotics in Healthcare Services: Applications in Silver Care. (2023). PubMed. Retrieved from https://pubmed.ncbi.nlm.nih.gov/40427894/
• Artificial Intelligence-Based Clinical Decision Support Systems in Geriatrics: An Ethical Analysis. (2023). PubMed. Retrieved from https://pubmed.ncbi.nlm.nih.gov/37453451/
• Artificial Intelligence in Mental Health. (2025). In Wikipedia. Retrieved from https://en.wikipedia.org/wiki/Artificial_intelligence_in_mental_health
• Applications of Artificial Intelligence. (2025). In Wikipedia. Retrieved from https://en.wikipedia.org/wiki/Applications_of_artificial_intelligence
• AI Ethics and Elderly Care. (n.d.). MEEGLE. Retrieved from https://www.meegle.com/en_us/topics/ai-ethics/ai-ethics-and-elderly-care
• AI and Ethical Robotics Set to Transform Elder Care Services. (2023). FTC Publications. Retrieved from https://news.ftcpublications.com/core/ai-and-ethical-robotics-set-to-transform-elder-care-services/
• Cambria, E., et al. (2017). ‘SenticNet 5: A Tool for Fine-grained Sentiment Analysis’. Proceedings of the 26th International Joint Conference on Artificial Intelligence (IJCAI), 5740-5742.
• CareYaya Health Technologies. (2024). QuikTok: An AI Companion Service for Older Adults. Retrieved from https://en.wikipedia.org/wiki/CareYaya_Health_Technologies
• Donaldson, G. C., et al. (2019). ‘Long lie times among older adults: Risk factors and health implications’. Geriatric Nursing, 40(6), 665-669.
• Eldercare Workforce Alliance. (n.d.). About Us. Retrieved from https://en.wikipedia.org/wiki/Eldercare_Workforce_Alliance
• Ethics of Artificial Intelligence. (2025). In Wikipedia. Retrieved from https://en.wikipedia.org/wiki/Ethics_of_artificial_intelligence
• Holt-Lunstad, J. (2017). ‘The potential public health relevance of social isolation and loneliness: Prevalence, epidemiology, and health consequences’. Public Health Reports, 132(5), 570-572.
• Lee, S. Y., et al. (2018). ‘Smart medication adherence system using internet of things technology’. Sensors, 18(12), 4383.
• Legal and Ethical Considerations of Artificial Intelligence for Residents in Post-Acute and Long-Term Care. (2023). PubMed. Retrieved from https://pubmed.ncbi.nlm.nih.gov/38909630/
• Liu, Y., et al. (2019). ‘Deep learning in Alzheimer’s disease: Diagnostic, prognostic, and therapeutic applications’. Trends in Neuroscience, 42(3), 169-179.
• Mubashir, M., et al. (2013). ‘A survey of fall detection systems based on wearable sensors’. Sensors, 13(9), 12104-12122.
• Navigating Ethical Considerations in the Use of Artificial Intelligence for Patient Care: A Systematic Review. (2023). PubMed. Retrieved from https://pubmed.ncbi.nlm.nih.gov/39545614/
• Pepper Robot. (n.d.). In Wikipedia. Retrieved from https://en.wikipedia.org/wiki/Pepper_(robot)
• Rashidi, P., & Mihailidis, A. (2013). ‘A survey on smart homes for elderly care’. IEEE Transactions on Systems, Man, and Cybernetics, Part C (Applications and Reviews), 43(5), 637-648.
• Rizzo, A. A., & Koenig, S. T. (2017). ‘Virtual reality for psychological assessments and interventions’. Annual Review of Clinical Psychology, 13, 109-130.
• Rougier, C., et al. (2011). ‘Fall detection from video: A survey’. Image and Vision Computing, 29(11), 740-745.
• Shibata, T. (2012). ‘Therapeutic seal robot PARO: A review’. Proceedings of the IEEE, 100(8), 2542-2547.
• Sutton, R. S., & Barto, A. G. (2018). Reinforcement Learning: An Introduction. MIT press.
• Top AI Trends in Senior Care. (n.d.). TMA Solutions. Retrieved from https://www.tmasolutions.com/insights/top-ai-trends-in-senior-care
• Topol, E. J. (2019). ‘High-performance medicine: the convergence of human and artificial intelligence’. Nature Medicine, 25(1), 44-56.
• United Nations, Department of Economic and Social Affairs, Population Division. (2020). World Population Ageing 2020 Highlights: Living arrangements of older persons. ST/ESA/SER.A/451.
• World Health Organization. (2021). Decade of Healthy Ageing: Baseline Report. Geneva: World Health Organization.