The power and perils of large language models in haematology

Large language models (LLMs) are a transformative technology poised to fundamentally change multiple fields including haematology. Here, we review the history of large language model development, describe their current capabilities and identify opportunities in haematology poised to benefit from their judicious application. We conclude by suggesting a framework for the safe and ethical development of LLM-derived tools in a way that minimizes risk and maximizes clinical benefit for haematology physicians and patients.