Duke Leads National Push for Trustworthy AI in Aging Care

Duke University will lead a five-year, nearly $25 million national initiative to build the infrastructure needed to evaluate and implement artificial intelligence tools for older adults and people living with dementia. The Duke Artificial Intelligence and Technology Collaboratory for Aging: Governance and Evaluation, known as Duke AITC-AIGE, is one of five centers selected for funding through the National Institute on Aging’s Artificial Intelligence and Technology Collaboratories program. The center is intended to create a shared pathway for moving promising AI systems from research prototypes toward responsible use in clinical and community settings.

The announcement describes a research and infrastructure program rather than the results of a clinical trial. Its central premise is that many AI algorithms have been proposed for aging and dementia care, but relatively few have undergone the rigorous evaluation, safe implementation and continuous monitoring required for routine practice. Duke AITC-AIGE will support pilot grants, data science resources, implementation assistance, governance tools and methods for assessing whether AI systems improve real-world outcomes for older adults and people with Alzheimer’s disease and related dementias.

Funding will provide about $5 million annually to the center, including approximately $4 million each year for research projects across the country. At least 80% of that research funding will be reserved for Alzheimer’s disease and related dementias, with the remainder supporting broader aging research. The nationwide program is designed to extend AI expertise beyond major academic medical centers and make support accessible to investigators at smaller institutions, community organizations, patients and caregivers who may have ideas for useful tools but lack the infrastructure required to build and test them.

The center is housed within the Duke Clinical and Translational Science Institute and is organized around three components. Juliessa Pavon, an associate professor in Duke’s Division of Geriatrics, leads the Administrative Core. Rishi Kamaleswaran, an associate professor of surgery and anesthesiology, leads the Clinical Implementation and Data Science Core. Sheng Luo, a professor of biostatistics and bioinformatics, leads the Research Innovation Core. Duke said the collaboration reflects the university’s capabilities in geriatrics, dementia research, translational science and responsible AI oversight.

Potential applications include detecting early cognitive decline by analyzing speech, behavior or data from wearable devices; identifying risks such as falls or wandering; supporting medication optimization and deprescribing; and helping caregivers monitor systems or access personalized guidance. These uses remain possibilities that require evaluation rather than established clinical benefits. The center’s role will include helping researchers determine which systems work in real-world settings, whether they improve outcomes and quality of life, and where their performance is insufficient or unsafe.

Transparency is a major focus of the initiative. Researchers and organizations will be able to develop and adapt model cards, standardized documents that summarize a tool’s intended purpose, training data, test performance, areas of strong performance and settings in which it may be less reliable. The cards can also document potential biases, limitations and appropriate uses. Pavon compared the idea to the nutritional information on food packaging, saying that clinicians and health systems should have a clear understanding of an AI tool’s risks and benefits before using it in a particular situation.

The Research Innovation Core will develop statistical and AI methods to identify performance limitations, investigate their causes and improve the reliability and generalizability of aging-related systems. Its work is particularly important because an AI model that performs well in a controlled research dataset may behave differently when applied to a new hospital, community or patient population. The press release does not report measured improvements from these methods; instead, it presents their development and testing as the center’s planned contribution to producing AI systems that can be assessed more systematically.

Duke will also develop the NeuroAI secure analytics and evaluation platform, which is intended to bring together medical records, brain images, laboratory results and genetic information. The platform will provide shared data and standardized methods for studying aging and dementia, developing AI tools and evaluating them. Duke said this common infrastructure could reduce the time, cost and specialized technical knowledge required to develop and test systems, making it easier for community researchers, smaller institutions, patients and caregivers to participate in innovation rather than leaving that work concentrated at large academic centers.

Beyond technical development, the center plans to provide national pilot funding and expertise in AI evaluation, data science, implementation, ethics, governance and stakeholder engagement. Its resources are intended to be reusable by investigators and organizations that are not part of a major academic AI center. Safety monitoring, implementation tools and standardized evaluation methods will sit alongside the research portfolio, reflecting the program’s emphasis that clinical adoption requires evidence about performance, oversight and continued operation in real environments.

The five-year goal is not simply to produce more algorithms, Pavon said, but to establish a clearer and reproducible route from innovation to safe, meaningful benefit. Success after five years would mean that researchers and health systems have better ways to determine which AI tools should enter practice, how they should be monitored and whether they support independence, care delivery, caregiver support and quality of life for older adults. The initiative therefore joins technical AI research with evaluation and governance infrastructure intended to address a central obstacle in translating aging-related AI from promising prototypes into dependable health tools.

Subject of Research: A national research infrastructure initiative for trustworthy artificial intelligence in aging and dementia care.

Article Title: Duke launches national hub to advance trustworthy AI for aging and dementia care

Article References: Duke launches national hub to advance trustworthy AI for aging and dementia care. Original publication (Bibliographic details unavailable.)

Image Credits: AI Generated

DOI: Not provided

Keywords: Duke University, trustworthy AI, aging care, dementia, Alzheimer’s disease, AI evaluation, clinical implementation, responsible AI, NeuroAI, National Institute on Aging, Duke, launches