Researchers from the Center have contributed to a newly published article in The Gerontologist calling for greater linguistic equity in the development of speech-based technologies for older adults. The article, titled “Equitable speech technologies for older adults: a call to prioritize low-resource languages,” was published on December 29, 2025, as part of a Gerontological Society of America (GSA) special issue on Artificial Intelligence and Aging.
Authored by Catherine Diaz-Asper, PhD, Mahederemariam Dagne, MS, and Patricia C. Heyn, PhD, FGSA, FACRM, the paper examines how recent advances in artificial intelligence can support healthy aging through non-invasive, scalable tools that promote social engagement, assist with daily living, and aid in disease detection.
Despite this promise, the authors note that most speech-based AI models are trained almost exclusively on English-language datasets, excluding over one billion older adults worldwide who speak non-English, low-resource languages. This linguistic imbalance limits access to emerging technologies, contributes to delayed diagnoses, and exacerbates existing health care inequities among aging populations.
The article emphasizes the urgent need to curate speech datasets in low-resource languages, guided by principles of community agency, equitable benefit sharing, and long-term sustainability. The authors argue that inclusive data practices are essential to ensuring speech technologies serve diverse older adults and advance equity in health care innovation.
