Healthy ageing is not simply the absence of disease. It is the capacity to preserve function, autonomy, social participation and quality of life for as long as possible. From this perspective, artificial intelligence (AI) can become a concrete ally, provided it is not presented as a futuristic substitute for care but as a responsible tool to strengthen prevention, personalisation and continuity of support across the life course. The ageing process is highly heterogeneous. Two people of the same chronological age may have markedly different biological profiles, functional reserves, nutritional needs, cognitive trajectories and exposure histories. This is precisely where AI can make a difference: by integrating information that is often fragmented across clinical records, laboratory data, lifestyle assessments, digital devices, environmental indicators and patient-reported outcomes. When appropriately validated, AI models can help identify early signs of vulnerability before they become clinically evident.
Frailty prevention is one of the most promising fields. Frailty is a dynamic condition that increases the risk of falls, disability, hospitalisation and loss of independence, but it can be delayed or partially reversed when detected early. Recent reviews show that machine learning models using real-world data, including structured clinical variables and information extracted from clinical notes, can predict frailty with moderate to high performance. However, the same literature also highlights a crucial point: these tools still require external validation, transparent reporting and evaluation within real clinical pathways before they can be considered ready for routine use.
AI may also support the measurement of biological ageing. New “ageing clocks” based on proteomic, metabolomic, epigenetic or multimodal data are helping researchers understand why some individuals accumulate age-related risk more quickly than others. A large Nature Medicine study published in 2024 showed that a proteomic ageing clock was associated with multimorbidity, mortality risk and physical and cognitive function. These models do not tell us how “old” a person is in a deterministic way; rather, they may help stratify risk and identify windows for targeted prevention. In the future, they could support clinical research into lifestyle medicine, nutrition, physical activity and inflammation as modifiable drivers of healthy ageing.
Nutrition is another area where AI can have a practical impact. Older adults are particularly vulnerable to malnutrition, sarcopenia, metabolic disorders and inappropriate dietary patterns. AI-based systems can integrate dietary intake, body composition, biochemical markers, medications, preferences and behavioural data to support more personalised nutritional counselling. This does not mean replacing dietitians or clinicians. On the contrary, AI can reduce repetitive tasks, improve monitoring between visits and help professionals identify when a person is deviating from an agreed nutritional or lifestyle plan.
The same logic applies to movement and falls prevention. Wearable sensors, smartphones and ambient technologies can generate continuous data on gait, balance, activity levels and sleep. AI can transform this data into indicators of functional decline or fall risk. This could be particularly valuable for older adults living at home, where conventional assessments are episodic and may miss subtle changes. Yet accuracy in controlled settings is not enough. For these technologies to be useful, they must work in real-world settings, be acceptable to older adults, respect privacy, and be integrated into care services that can respond when an alert is generated.
Cognitive health is another priority. AI is increasingly used to analyse neuroimaging, speech, behaviour, sleep, mobility and electronic health record patterns to support earlier detection of cognitive impairment or dementia-related changes. Early detection may enable timely clinical assessment, lifestyle interventions, caregiver support and better planning. At the same time, cognitive health is a sensitive domain: false alarms, stigma, data misuse and over-reliance on automated outputs must be carefully avoided.
Notably, AI can concretely support healthy ageing only if embedded within a trustworthy medical and public health framework. Healthy ageing is multidimensional: it includes nutrition, physical activity, sleep, mental health, social relationships, environmental exposures, access to care and health literacy. AI can help connect these dimensions, but it must remain explainable, inclusive and clinically meaningful.
This also requires attention to equity. The World Health Organisation has warned that AI systems can reproduce or amplify ageism if older adults are under-represented in datasets, excluded from design processes, or treated as a homogeneous population. Digital health tools must therefore be co-designed with older adults, caregivers, and healthcare professionals. Accessibility, language, usability, affordability, and digital literacy are not secondary details: they determine whether innovation reduces or increases inequalities.
A responsible AI approach to healthy ageing should adhere to clear principles. First, models should be trained and tested on diverse populations. Second, predictions should be interpretable and used to support, not replace, professional judgement. Third, data governance must protect privacy and autonomy. Fourth, benefits should be evaluated using outcomes that matter to older people: preserved function, reduced disability, better quality of life, fewer unnecessary hospitalisations, and greater participation in society.
In this sense, AI is not a magic solution for ageing. It is an enabling technology. Its value depends on the quality of the data, the robustness of validation, the ethics of implementation and the human intelligence with which it is used. If developed responsibly, AI can help move healthy ageing from a reactive model, focused on treating disease after it appears, to a proactive model, focused on preserving intrinsic capacity, personalising prevention and supporting longer lives lived in better health.
References
- World Health Organization. Ethics and governance of artificial intelligence for health: guidance on large multi-modal models. WHO, 2024.
- World Health Organization. Ageism in artificial intelligence for health. WHO Policy Brief, 2022.
- Bai C, Zhang Y, Zhao X, et al. Advances of artificial intelligence in predicting frailty using real-world data: a scoping review. Ageing Research Reviews. 2024; doi: 10.1016/j.arr.2024.102529.
- Ou Y, Jiang D, Li P, Zhang W, Zhou Y, Chen Y, Yin X. Prediction of frailty in community older adults based on machine learning: a systematic review and meta-analysis. Frontiers in Public Health. 2026; doi: 10.3389/fpubh.2025.1667792.
- Argentieri MA, Amin N, Nevado-Holgado AJ, et al. Proteomic ageing clock predicts mortality and risk of common age-related diseases in diverse populations. Nature Medicine. 2024. doi: 10.1038/s41591-024-03164-7.
- Lu JK, et al. Digital biomarkers of ageing for monitoring physiological function and health trajectories. The Lancet Healthy Longevity. 2025. doi: 10.1016/j.lanhl.2025.100725
- Agrawal K, et al. Artificial intelligence in personalized nutrition and food manufacturing: a comprehensive review of methods, applications, and future directions. Frontiers in Nutrition. 2025; doi: 10.3389/fnut.2025.1636980.
- Fernandes S, Rosselet Amoussou J, Gomes da Rocha C, Perruchoud E, von Gunten A, Mabire C, Verloo H. Using artificial intelligence–based technologies for the early detection of behavioral and psychological symptoms of dementia: scoping review. JMIR Aging. 2025; doi: 10.2196/76074.
