JMIR Publications

ABSTRACT

Background:

Embodied intelligence—artificial intelligence systems integrated into physical or virtual bodies that interact with users and environments—has been increasingly applied in healthcare. Despite rapid growth in empirical research, existing evidence remains fragmented, with inconsistent terminology and limited synthesis of intervention types, functional roles, target populations, and application settings. This fragmentation constrains the translation of embodied intelligence research into real-world healthcare practice.

Objective:

This scoping review aimed to systematically map the applications of embodied intelligence in healthcare, with a focus on classifying forms of embodied intelligence, identifying core functional domains, summarizing target populations and application settings, and synthesizing evidence on acceptability and effectiveness.

Methods:

This review followed the Arksey and O’Malley framework and was reported in accordance with the PRISMA-ScR and PRISMA-S guidelines. A comprehensive search of seven electronic databases (PubMed, Web of Science, Cochrane Library, MEDLINE, APA PsycNet, CQVIP, and CNKI) was conducted from database inception to December 2025, supplemented by grey literature searches and reference list screening. Primary empirical studies of any design published in English or Chinese were eligible if they examined embodied intelligence (ie, AI with a physical or virtual form enabling situated interaction) in any healthcare setting; studies were excluded if the technology lacked an embodied form (eg, purely text-based chatbots), were non-primary research, or full texts were unavailable. Two reviewers independently screened studies and extracted data using a pre-tested standardized data charting form. Descriptive statistics and thematic synthesis were applied. As this was a scoping review and the evidence base was emerging, a formal critical appraisal of study quality (risk of bias assessment) was not conducted.

Results:

A total of 83 studies were included from 3,797 screened records. Five forms of embodied intelligence were identified: virtual humanoid agents (32/83, 38.6%), physical humanoid robots (32/83, 38.6%), virtual animal-shaped agents (1/83, 1.2%), physical animal robots (13/83, 15.7%), and mechanical robots (5/83, 6.0%). Applications clustered into three functional domains: health management and health education (40/83, 48.2%), mental health promotion (37/83, 44.6%), and physiological health promotion (6/83, 7.2%). Older adults were the most frequently targeted population (45/83, 54.3%). Interventions were primarily implemented in home settings (27/83, 32.1%) and care home settings (22/83, 25.9%), followed by lab environments (23/83, 27.2%) and hospital settings (11/83, 14.8%). Twenty-two randomized controlled trials consistently reported beneficial effects on health behaviors, mental health outcomes, or cognitive function, although outcome measures were heterogeneous. Twelve studies assessed acceptability and generally reported high user acceptance, particularly when interaction quality and anthropomorphic design were optimized.

Conclusions:

This scoping review provides the first comprehensive, evidence-based synthesis of embodied intelligence in healthcare using a unified classification of forms, functional domains, populations, and application settings. The findings indicate that embodied intelligence is most mature in health management and health education and mental health promotion, with growing real-world deployment in home and care home settings. By consolidating fragmented evidence and standardizing terminology, this review offers a practical foundation for clinicians, nurses, designers, and policymakers to support the implementation of embodied intelligence in routine healthcare. Evidence is limited by heterogeneous outcome measures, the presence of many lab-based evaluations, and the absence of formal quality appraisal, underscoring the need for standardized outcome measures, rigorous randomized controlled trials, and longitudinal evaluations to enable scalable and ethically grounded real-world adoption. Clinical Trial: None