Governing Generative AI for Healthy Ageing: A Normative Conceptual Framework for Societal Alignment, Epistemic...

Large language models (LLMs) and generative AI are rapidly being integrated into healthy ageing initiatives for tasks ranging from companionship and cognitive support to personalised health advice and reduction in social isolation among older adults. Current ethical discussions predominantly address bias, privacy, and accuracy, leaving unresolved three critical governance questions: How do LLM sentiments towards transformative technologies diverge from human values in ageing contexts? What epistemic status do LLM outputs hold when applied to geriatric care? When is trust in those outputs justified for older adults? And who bears responsibility when AI-informed decisions affect functional ability or well-being?

The framework was developed through normative conceptual analysis, synthesizing philosophical principles of medical knowledge and trust, ethical theories of responsibility, empirical evidence on LLM sentiment divergence, digital ageism, and applications of AI in geriatric care (structured searches in PubMed, PhilPapers, and relevant databases, January 2020-March 2026).

The integrated framework produces (i) adaptation of SAIA for multidimensional evaluation of human-machine value convergence specific to healthy ageing values (functional ability, autonomy, dignity, equity); (ii) a four-tier classification of LLM outputs tailored to geriatric scenarios; (iii) conditions for warranted trust calibrated to age-related vulnerabilities such as cognitive decline and digital divide; and (iv) responsibility allocation via RACI models with testable hypotheses linking governance design to trust calibration and patient safety outcomes.

Without explicit societal alignment and epistemic governance, generative AI risks reinforcing benevolent ageism, automation bias, and responsibility gaps in healthy ageing. The 2025-2027 period offers a decisive window to shape institutional norms that place functional capacity, human dignity, and value convergence at the centre of AI deployment in geriatric care.