This overview of ethical aspects of artificial intelligence (AI) in healthcare is organized according to major types of medical activities. It discusses important ethical problems in (1) diagnostic uses of AI such as image interpretation and text interpretation, (2) the use of AI as decision support or even as decision-maker, (3) its use for treatment, for instance in robotic surgery, exoskeletons, virtual psychotherapy and therapeutic robots (care robots), and (4) the use of AI for healthcare documentation and communication. Summing up, we emphasize two major challenges in the introduction of AI that should concern not only healthcare professionals but also higher level decision-makers such as politicians and hospital directors. The first of these is the need to retain human contacts between patients and healthcare, in particular contacts that are conducive to patient health and well-being. The second is the need to ensure that the introduction of AI in healthcare does not create or exacerbate injustices in healthcare. This can happen if prevalent human biases are encoded in AI tools. It can also occur if people with less money have to use inexpensive AI tools such as virtual psychotherapists whereas the well-to-do have access to a human psychotherapist. On the other hand, AI tools can be used to provide better healthcare to all who need it, and they can also be programmed to avoid biases and other mistakes that humans are prone to. The article concludes with brief summaries of the articles in the special issue on AI in healthcare that it introduces.
Read full abstract