Using Agentic AI to Review GO Standard Annotations Abstract This presentation introduces an agentic AI approach to systematically review and improve Gene Ontology (GO) annotations at scale. Traditional GO curation requires expert biocurators to manually evaluate evidence for thousands of gene-function associations—a time-intensive process that struggles to keep pace with the growing volume of biological literature. The ai-gene-review system (https://github.com/ai4curation/ai-gene-review) addresses this challenge by deploying AI agents equipped with specialized bioinformatics tools. The workflow combines deep literature research with real-time access to GO databases (QuickGO), protein databases (UniProt), pathway resources (Reactome), ontology services (OLS), and PubMed. Agents can also write and execute ad-hoc Python code for sequence analysis when needed. For each gene, the system reviews every existing annotation and assigns one of six actions: ACCEPT (retain as core function), MODIFY (propose better terms), REMOVE (insufficient evidence), MARK_AS_OVER_ANNOTATED (too generic or indirect), KEEP_AS_NON_CORE (valid but peripheral), or NEW (propose missing annotations). Each decision includes a detailed rationale synthesizing multiple evidence sources. Key capabilities demonstrated include: - Detecting over-annotations inferred from indirect evidence - Identifying outdated annotations superseded by new literature - Proposing more precise GO terms when annotations are too general - Discovering missing annotations from recent publications - Performing sequence-based analyses to validate functional assignments Initial evaluation across diverse genes (human, yeast, bacteria, phage) shows the system can identify problematic annotations that would require expert attention while providing transparent, traceable reasoning. Limitations include occasional hallucination risks and challenges with complex isoform-specific functions.