Abstract
Artificial intelligence (AI) has shown great potential in the field of hepatology, and it is gradually revolutionizing novel research for liver diseases. However, its full integration into real-world diagnostic and treatment processes still faces significant challenges. This article analyzed the application progress of AI in key areas such as big data analysis, translational research, and imaging and pathological interpretation. Physicians need to integrate massive amounts of multimodal data (medical history, physical signs, laboratory data, imaging, and pathology) to make informed decisions in clinical practice. Although AI has gradually shown a preliminary auxiliary application role in certain liver disease diagnosis and treatment scenarios, current machine learning and deep learning technologies are far from being able to effectively support clinical decision-making in real-world settings. The core pain points are as follows: firstly, AI tools are not yet mature enough to truly independently and reliably assist in complex clinical evaluation; secondly, even if assistance is provided, the ultimate responsibility for medical decisions still clearly belongs to the physician, failing to reduce their workload effectively; finally, due to the different specialties and data characteristics of various medical centers, individually trained AI models are often relatively independent and personalized, fragmented, and difficult to integrate into a widely applicable and scalable multi-center general model, limiting their universal value. In the future, enhancing clinicians' understanding of AI and building multidisciplinary collaborative networks will be crucial to accelerating the development of AI-driven decision support tools for specific liver diseases. At the same time, ethical, legal, and talent development challenges, and others must be carefully addressed to bridge the gap and safely and eppectively apply AI in clinical practice.