Digital technologies and artificial intelligence in eating disorders: A scoping review of prevention, screening, and treatment approaches

Eating disorders (EDs) are prevalent and often underdiagnosed, and digital interventions may reduce barriers to prevention, early detection, and care. This scoping review mapped research published in 2015–2025 on digital health and artificial intelligence (AI) applications for ED prevention, screening/assessment, prediction/monitoring, and treatment. Searches were run on 29 December 2025 via the University of Catania Primo VE discovery service, covering collections including Scopus, PubMed, PsycArticles, and others. Records were deduplicated and screened using a human-in-the-loop workflow with Large Language Model prioritization used only to order citations for manual screening; study data were charted with a standardized form and synthesized descriptively. The search retrieved 9390 records; 1845 duplicates were removed, leaving 7545 unique records. Screening focused on the top 500 prioritized citations; 191 full texts were assessed and 40 studies were included. Included evidence split evenly between digital technologies (n = 20; mainly web/app cognitive-behavioral programs, screening/triage platforms, and virtual reality interventions) and AI (n = 20; mainly machine learning and natural language processing approaches for risk detection and prediction, plus emerging chatbot-based supports). Evidence was heterogeneous and often early-stage, with limited external validation and uneven reporting of safety and implementation. Future work should emphasize workflow-integrated evaluations, engagement and equity, and governance frameworks for automated tools.