Ethical barriers to artificial intelligence in the national health service, United Kingdom of Great Britain and Northern Ireland

Artificial intelligence, the ability of systems to replicate human behaviour in an intelligent manner, shows promise in the United Kingdom of Great Britain and Northern Ireland’s National Health Service (NHS), which provides freeat-the-point-of-service health care via a national insurance scheme (Fig. 1). Recent advancements in artificial intelligence have created sophisticated software programmes that could revolutionize the NHS. Breakthroughs in machine learning, more notably deep learning (Box 1), have led to algorithms capable of performing diagnostic skills equivalent to those of doctors, automating administrative tasks and assisting in complex treatment management. As a result of the vast linkable data that the NHS holds on all citizens throughout their lives, the service could have a leading role in taking forward artificial intelligence development for health care;1 however, its use remains limited, with little overarching policy guiding its development and application. In 2018, the government of the United Kingdom published a code of conduct outlining expectations for artificial intelligence development in the NHS, covering aspects such as the appropriate handling of data, the need for algorithmic transparency and accountability.2 The code states that, in combination with the conformité européenne (CE) mark certification, health research and relevant regulatory approvals, it should provide an overall policy and structure for the creation of safe and effective artificial intelligence. The code, however, is only in its initial consultation stage. This paper discusses the issues highlighted within the code of conduct and the ethical challenges associated with addressing them to successfully integrate artificial intelligence within the NHS. Patient data in training algorithms