From In Silico Hypothesis to Validation: The Role of Real-World Evidence in the Preliminary Verification of AI-Generated Drug-Repositioning Candidates: A Comprehensive Review

Background/Objectives: Drug repositioning has emerged as a promising strategy to address the innovation crisis in pharmaceutical development. While artificial intelligence enables efficient in silico hypothesis generation, clinical translation remains challenging. This study aims to evaluate the role of Real-World Evidence (RWE) in validating AI-generated drug-repositioning candidates. Methods: A comprehensive literature review was conducted in PubMed using a predefined search strategy integrating drug repositioning, artificial intelligence, and real-world data. After multi-stage screening, 22 original research articles were included for analysis. Results: Network-based algorithms and natural language processing dominated AI-driven hypothesis generation. Validation using Electronic Health Records and insurance databases enabled retrospective assessment of drug efficacy across large populations. Successful applications were identified in neurodegenerative, metabolic, infectious, autoimmune, and psychiatric diseases. Conclusions: The integration of AI-based analytics with RWE provides a promising framework for the preliminary verification of computational predictions, potentially informing the translational pathway toward clinical practice. However, the effectiveness of this approach remains dependent on data quality and the specific therapeutic context, requiring further standardization of clinical data.