e13658 Background: Scientific articles are being retracted at higher rates for many reasons. Retractions or corrections may be unknown to authors citing these materials as primary sources, unwittingly perpetuating incorrect information. Our goal was to assess the utility of a scholarly artificial intelligence (AI) tool to identify errata to sources cited in NCCN Clinical Practice Guidelines in Oncology (NCCN Guidelines) for gynecologic cancers and to evaluate the clinical importance of identified errata. Methods: scite (scite Inc., Brooklyn, NY) is an AI tool that analyzes citations to generate a list of linked publications, display the context for each citation, and flag any published errata or retractions for each source. Using scite with references from NCCN Guidelines as of June 2023, we identified corrections to primary sources. Each author categorized errata as having major, minor, or no clinical relevance, with resolution of conflicts through discussion until consensus. To check whether scite missed errata, we assessed a random sample (20%) of unflagged references. Results: The guidelines cited 2,364 primary sources, and scite successfully linked to 1,984 (84%) of them . scite flagged 50 references (2.5%) as having corrections; all 50 did have errata (100% accuracy). scite identified 56 total errata, which were not always apparent on journal websites or PubMed. After review, 31 errors (55%) were classified as clinically relevant; 71% of those errors had major clinical importance and 29% were considered minor. Twenty errors (65%) had relevance to the guideline statements in which the source was cited, but only 1 (3%) error had potential to change current guideline text. Among the sample of sources scite did not flag, we identified none with published errata. Conclusions: We demonstrated use of an AI tool to accurately identify sources in clinical guidelines with published corrections. Sources with errata were uncommon, and our analysis did not reveal perpetuation of major errors in NCCN Guidelines. However, the proportion of corrections with major clinical relevance underscores the importance of awareness of errata when writing clinical guidelines or reviews. Compared to manual review of each source, AI tools can improve speed and accuracy in finding important errors. [Table: see text]