A new position statement from the American Geriatrics Society (AGS) offers one of the first comprehensive frameworks for the responsible use of generative artificial intelligence (GenAI) in the care of older adults. Published in the Journal of the American Geriatrics Society, the statement outlines opportunities and risks associated with rapidly expanding AI technologies and provides recommendations for clinicians, health systems, developers, policymakers, and care partners

The authors note that GenAI tools, including large language models, are increasingly being used in healthcare for clinical documentation, decision support, patient education, and administrative tasks. While these tools may help reduce clinician burden and improve efficiency, they also carry risks such as inaccurate information, algorithmic bias, privacy concerns, and overreliance on automated recommendations.

According to the AGS, older adults may be particularly vulnerable to these risks because important aspects of geriatric care, such as cognitive assessment, goals-of-care discussions, polypharmacy management, and functional status evaluation, may be underrepresented in AI training datasets. As a result, AI-generated recommendations may not always reflect an individual’s priorities, health conditions, or care preferences.

The position statement emphasizes that GenAI should augment, rather than replace, clinical judgment and person-centered care. Key recommendations include maintaining clinician oversight of AI-generated outputs, ensuring transparency when AI tools are used in patient care, validating AI systems in older adult populations, and establishing governance processes to monitor for unintended harms and inequities.

For Alzheimer’s disease and related dementias (ADRD) researchers, the guidance highlights the importance of developing and evaluating AI tools that account for cognitive impairment, decision-making capacity, and the needs of care partners. The authors also call for meaningful involvement of older adults and clinicians in the design, implementation, and oversight of AI technologies.

The Duke/UNC ADRC community is well represented in this national effort. Dr. Juliessa M. Pavon of Duke University School of Medicine was among the interdisciplinary experts who developed the AGS position statement. As AI tools become increasingly integrated into healthcare and research, contributions from aging and dementia researchers help ensure these technologies are developed and evaluated with the needs of older adults, care partners, and clinicians in mind.

Reference: Bharija A, Khan A, Pavon JM, et al. Generative AI in the Care of Older Adults: A Position Statement From the American Geriatrics Society. Journal of the American Geriatrics Society. 2026.