Oniani et al. delve into the ethics of generative AI in both the military and healthcare system. First, the authors discuss the potentially dangerous outcomes of AI in healthcare, including algorithmic bias, hallucinations, poor commonsense reasoning, and lack of generally agreed model evaluation metrics. Next, the paper talks about optimal trust levels for AI technology, particularly in the healthcare system – too much trust can lead to overuse, too little trust can lead to underuse. The authors then propose a framework to guide the ethical implementation of generative AI into the healthcare system, referred to as the “GREAT PLEA” – Governability, Reliability, Equity, Accountability, Traceability, Privacy, Lawfulness, Empathy, and Autonomy. Primarily this article focuses on the ethics dilemma imposed by the integration of generative AI into the healthcare system, using frameworks that have been useful in the military to guide the theory.
