AI-powered cognitive remediation for schizophrenia: a psychologist's lens on evidence and scalability in India

Abstract

Background Schizophrenia spectrum disorders (SSDs) are marked by cognitive deficits, functional impairment, and notable treatment gap in India, with lifetime prevalence of 1.41%, current prevalence as 0.42%, and the gap of 72%, added by a severe workforce gap, that is, 0.75 psychiatrists per 100,000 and 0.29 clinical psychologists per 100,000. Scalable tools are required by the psychologists that can fill the substantial urban-rural gap. Aim To evaluate how effective, feasible, and clinically applicable AI-powered cognitive remediation is, emphasizing India/LMIC contexts. Methods PRISMA-guided umbrella review with narrative synthesis of records published between 2020–2026, identified 48 studies (14 randomized controlled trials (RCTs), 22 reviews, 8 pilots, and 4 bibliometric studies) across PubMed, Scopus, Web of Science. Cognition, functioning, and feasibility, retention, and usability were included as outcome measures. Results AI-based AI-powered cognitive remediation showed generally favorable effects. Reported effect sizes ranged from moderate to large across different interventions and outcomes. VR-based social-cognition training showed largest benefits and smartphone-based programs showed promise for working memory and engagement. Hybrid or monitored formats showed generally better retention and usability than fully self-directed interventions. However, the evidence remained heterogeneous, and many findings came from small studies, feasibility work, or reviews of mixed quality. Conclusion AI-powered CRT appears to be a promising model for schizophrenia rehabilitation when delivered through psychologist-monitored hybrid models. Such approaches may improve access, support rural and semi-rural service delivery, and help reduce the treatment gap through scalable, culturally adaptable, and technology-enabled care, especially in rural India.