Elevating Nursing Care and Health Outcomes: An In-Depth Qualitative Synthesis Exploring the Transformative Impact of Assistive Technology (AI) on Individuals with Disabilities

Abstract

Artificial intelligence has been incorporated into assistive products intended to support mobility, communication, sensing, cognition and social participation among people with disabilities, and nursing staff increasingly mediate the assessment, introduction and everyday use of these products. Claims about transformative effects on health outcomes have outpaced the evidence assembled to support them. This critical narrative review examines what can defensibly be concluded about artificial intelligence-enabled assistive technology in the care of people with disabilities, with particular attention to the position of nursing. Peer-reviewed literature was identified through openly accessible scholarly indexes and citation searching, complemented by reports from recognised international institutions, and was appraised for design adequacy, outcome relevance, representativeness and consistency. Five arguments emerge. The field is conceptually unsettled, because adaptive assistive products, ambient monitoring systems and clinical prediction models are frequently grouped together despite differing in who acts on the output and who carries the resulting risk. Evidence of function-level improvement is strongest for locomotor and communication technologies, yet pooled estimates are modest, inconsistent across outcome domains and derived largely from small trials rated as providing low certainty. Outcome measurement is fragmented, with satisfaction and usability instruments predominating over participation and health endpoints, which limits comparison across studies and obscures whether functional gains translate into health benefit. Nursing appears in this literature mainly as an object of attitude surveys rather than as a determinant of implementation, despite evidence that staff interpretation, workload and relational judgement shape whether devices are used or abandoned. Benefit is conditional on access, data representation and governance, and each condition is unevenly distributed, with interventional evidence concentrated in high-income settings, older populations and dementia-related conditions. Priorities include pragmatic trials with participation and nursing-sensitive endpoints, harmonised outcome sets, disaggregated reporting of algorithmic performance across impairment subgroups, and evaluation of nurse-mediated implementation. The accessible evidence supports cautious, context-specific optimism rather than claims of transformation.