Avian influenza viruses (AIV) exhibit high mutation rates and are continuously introduced into South Korea via migratory wild birds, leading to recurrent outbreaks of highly pathogenic avian influenza (HPAI). This study aimed to identify genetic mutations associated with increased viral spread and zoonotic risk and to predict future genomic evolution using an AI-based approach.

A total of 453 HPAI-positive isolates collected from wild birds between 2020 and 2023 were analyzed, covering all eight gene segments (PB2, PB1, PA, HA, NP, NA, M, NS). Sequence alignment and variant calling were performed to identify mutation sites, and mutation frequencies were quantified across years. Sixteen significant mutations were identified (p < 0.001), among which eight hotspot mutations were strongly enriched during 2021–2022, coinciding with increased outbreak cases. These mutations are likely associated with enhanced polymerase activity and immune evasion.

Mutation frequency trends were modeled over time and extrapolated to predict 2024 genomic changes. Validation using independent datasets demonstrated high predictive accuracy, particularly for polymerase genes. These findings suggest that specific mutation markers may serve as early indicators of HPAI spread and zoonotic risk, contributing to improved surveillance and control strategies.