Ethics of artificial intelligence in global health: Explainability, algorithmic bias and trust

AI has the potential to disrupt and transform the way we deliver care globally. It is reputed to be able to improve the accuracy of diagnoses and treatments, and make the provision of services more efficient and effective. In surgery, AI systems could lead to more accurate diagnoses of health problems and help surgeons better care for their patients. In the context of lower-and-middle-income-countries (LMICs), where access to healthcare still remains a global problem, AI could facilitate access to healthcare professionals and services, even specialist services, for millions of people. The ability of AI to deliver on its promises, however, depends on successfully resolving the ethical and practical issues identified, including that of explainability and algorithmic bias. Even though such issues might appear as being merely practical or technical ones, their closer examination uncovers questions of value, fairness and trust. It should not be left to AI developers, being research institutions or global tech companies, to decide how to resolve these ethical questions. Particularly, relying only on the trustworthiness of companies and institutions to address ethical issues relating to justice, fairness and health equality would be unsuitable and unwise. The pathway to a fair, appropriate and relevant AI necessitates the development, and critically, successful implementation of national and international rules and regulations that define the parameters and set the boundaries of operation and engagement.