Duke launches AI center with nearly $25 million to advance aging and dementia care

Duke is building a national pathway for testing whether AI can actually improve care for older adults and people living with dementia. Announced October 12, its new center will receive nearly $25 million over five years from the National Institute on Aging to fund research and help move promising tools toward real-world use.

The catalyst, according to co-principal investigator Juliessa Pavon, is not a lack of algorithms. It is a shortage of AI solutions that have been rigorously evaluated, safely introduced into practice and continuously monitored—and shown to improve outcomes for older adults. The center’s response is to build the infrastructure and collaborations needed to take those steps.

The Duke Artificial Intelligence and Technology Collaboratory for Aging: Governance and Evaluation, or AITC-AIGE, is one of five centers selected under the institute’s collaboratories program. Housed within Duke’s Clinical and Translational Science Institute, it will combine research funding with data science expertise, implementation support and responsible-AI safeguards.

Research funding reaches beyond Duke

Of roughly $5 million in annual funding, about $4 million will go to research projects nationwide. At least 80% of that research funding will support Alzheimer’s disease and related dementias, with the remainder supporting broader aging research, according to co-principal investigator Sheng Luo. That allocation makes dementia the dominant research priority without restricting the center to it.

Pavon identified several potential applications: detecting early cognitive decline through speech, behavior and wearable-device data; identifying risks such as falls or wandering; and helping optimize medication use. Caregiver support is another area of interest. These are research possibilities the center aims to evaluate in real-world settings, rather than benefits already demonstrated by its launch.

We’re about improving lives, not just improving these algorithms

Juliessa Pavon, Duke AITC-AIGE co-principal investigator

Making the evidence easier to inspect

The planned infrastructure addresses both data access and scrutiny of the resulting tools. Through NeuroAI, a secure analytics and evaluation platform, researchers will be able to bring together medical records, brain images, laboratory results and genetic information. The aim is to use those different kinds of health data to study aging and dementia and develop and evaluate AI.

The center also plans evaluation methods, safety-monitoring approaches and implementation tools that other organizations can adopt and adapt. Its model cards—standardized summaries resembling nutrition labels—are intended to make an AI tool’s strengths and weaknesses understandable to clinicians, health systems, researchers and patients. Duke says those cards will explain:

• What a tool is designed to do and the data used to train it.
• How it performed during testing, where it works well and where it may be less reliable.
• Potential biases, limitations and appropriate uses, helping users judge whether it suits a particular situation.

Alongside that disclosure work, the Research Innovation Core will develop statistical and AI methods to identify performance limitations, understand their causes and improve reliability across settings. The planned work therefore includes both describing a tool’s weaknesses and investigating how to address them.

A pathway smaller institutions can use

A national pilot funding program is intended to give investigators access to evaluation, data science, ethics, governance and implementation expertise without belonging to a large academic AI center. Co-principal investigator Rishi Kamaleswaran said the effort should extend expertise to community researchers and even patients with ideas but insufficient resources to build a tool.

Pavon’s five-year measure of success is a clearer, reproducible pathway for researchers and health systems to turn AI innovation into safe, meaningful benefits for older adults. That sets an outcome-focused ambition for the center: not simply producing more tools, but establishing how promising ones can become useful in care.