Occupational Data Stewardship in the Era of Artificial Intelligence: A Call for Occupational Therapy Leadership.

Abstract

Artificial intelligence (AI) is rapidly reshaping occupational therapy, and the literature has increasingly addressed robotics, predictive models, assessment technologies, ethics, documentation, and workforce readiness. Comparatively little attention, however, has been given to whether the occupational data underlying AI systems adequately represent occupation or support occupation-centered practice. In this article, representative occupational data are defined as data that capture occupation as it is experienced in context rather than reducing it to impairment, task performance, or service use alone. I contend that occupational therapy practitioners have a professional responsibility for occupational data stewardship by shaping how occupational data are structured, interpreted, governed, and applied across AI-supported practice, education, and research. This work includes identifying occupation-related information, translating it into meaningful data, and critically evaluating how those data are used within AI systems. To advance occupational data stewardship, I propose four priorities: establishing minimum occupational data elements, protecting occupational data privacy, maintaining human-in-the-loop accountability, and implementing AI in ways that prioritize occupational justice and participation over efficiency.