Artificial intelligence for hepatobiliary and pancreatic cancer: Ethics, equity, and real-world integration

Hepatobiliary and pancreatic (HPB) cancers present a major challenge due to their late presentation, limited treatment options, and high mortality. Artificial intelligence (AI) has emerged as a promising tool in revolutionising cancer care, offering potential advances in early detection, and treatment planning. However, real-world implementation of AI remains limited by ethical, technical, and systemic challenges. This narrative review explores the evolving landscape of AI in HPB oncology, with a focus on ethical integration, healthcare equity, and clinical applicability. Key issues discussed include algorithmic bias, informed consent, model explainability, and disparities in access to data and AI-driven tools. Promising innovations such as federated learning and large language models are explored for their potential to decentralise model training and enhance multidisciplinary workflows. The review also highlights the integration of AI into surgical navigation systems and intraoperative decision-making, as well as its application to omics data analysis for biomarker discovery. Crucially, it underscores the need for transparent and interpretable systems, the need for prospective validation in diverse populations, and the risk of clinician de-skilling. As AI technologies evolve, their safe and equitable integration into HPB oncology will require robust governance, regulatory foresight, and sustained investment in clinician education and infrastructure. This review concludes that, while AI shows potential in transforming HPB cancer care, its ethical and inclusive implementation will ultimately determine its clinical impact.