Beyond prediction: The evolving role of artificial intelligence in genome editing

Since the discovery of the clustered regularly interspaced short palindromic repeat (CRISPR)-associated (Cas) system, genome-editing technologies have advanced rapidly, and the field has expanded substantially. In parallel, artificial intelligence (AI) has progressed at an unprecedented pace and is transforming genetic engineering. Early AI applications primarily addressed predictive tasks, including guide RNA (gRNA) activity, off-target effects, and editing outcomes for CRISPR-Cas nucleases, base editors, and prime editors. Recently, transformer-based protein language models (PLMs) and generative AI have expanded the role of AI beyond prediction to direct molecular design. Thus, AI-driven frameworks can now evolve genome-editing proteins, design molecular architectures, and generate de novo genome-editing enzymes. This review discusses how AI is reshaping genome engineering, from gRNA activity prediction to the design of next-generation editors.