Over the past decade, AI has been used to aid or even replace humans in many professional fields. There are now robots delivering groceries or working in assembling lines in factories, and there are AI assistants scheduling meetings or answering the phone line of customer services. Perhaps even more surprisingly, we have recently started admiring visual art produced by AI, and reading essays and poetry “written” by AI (Miller 2019), that is, composed by imitating or assembling human compositions. Very recently, the development of ChatGPT has shown how AI could have applications in education (Kung et al. 2023) the judicial system (Parikh et al. 2019) and the entertainment industry.1
One of the most promising areas of development for AI is healthcare (Mishra et al. 2021). AI has been used for a few years to assist in diagnosing patients and find the best treatments (see, for instance, IBM Watson). More recently, AI robots have been used to help surgeons perform brain surgery (Prabu et al. 2014).
AI is quickly becoming effective at performing several tasks in healthcare settings that we used to consider a human prerogative. In particular, AI seems to be better than humans at diagnosing some diseases because it can learn from vast datasets and recognise patterns better than we can (Loh 2018). It is therefore likely that areas of medicine mainly concerned with diagnosing will be integrated with AI sooner than others.
Demand for healthcare assistance is constantly increasing, but financial resources are limited. The prospect of using AI to provide substantial support with the administration of healthcare has made many hopeful about the possibility of bettering healthcare assistance worldwide and making it more cost-effective. Reports of its successful use in medicine have appeared in both scientific journals and popular magazines. For instance, some studies suggest that AI is better than human doctors at detecting conditions such as skin cancers (Esteva and Topol 2019) or diabetic retinopathy (Savoy 2020), and some predict that AI will replace human healthcare practitioners within a few decades2.
Assuming that AI can deliver on what it promises—a big assumption we take for granted in this article—AI could allow better care for a greater number of people. Yet the costs of switching from human to AI medical assistance cannot be dismissed.
One of the obvious costs associated with replacing a significant number of human doctors with AI is the dehumanization of healthcare. The human dimension of the therapist-patient relationship would surely be diminished. With it, features of human interactions that are typically considered a core aspect of healthcare provision, such as empathy and trust, risk being lost as well. Sometimes, the risk of dehumanizing healthcare by having machines instead of persons dealing with patients might be worth taking, for instance when the expected outcomes for the patient are significantly better. However, some areas of healthcare seem to require a human component that cannot be delegated to artificial intelligence. In particular, it seems unlikely that AI will ever be able to empathize with a patient, relate to their emotional state or provide the patient with the kind of connection that a human doctor can provide. Quite obviously, empathy is an eminently human dimension that it would be difficult, or perhaps conceptually impossible, to encode in an algorithm.
In some areas of healthcare, these factors might be almost irrelevant. One might care little about having a human connection with the entity (human or not) filling a cavity or performing surgery on a broken finger. But in areas such as psychiatry and mental health care, interaction with another human is likely to be irreplaceable, as it seems to be one of the main factors contributing to successful psychiatric diagnosis and treatment.
However, if using AI in psychiatry yields positive results for the patients, not only would the dehumanisation of psychiatry be a cost worth paying—especially if it reduces costs and improves access; in addition, the very nature of psychiatry as essentially grounded in the human connection between the therapist and the patient would be called into question. Even if there are reasons to be skeptical about the possibility of this radical change of perspective, it is worth looking at what the potential benefits might be. It is possible that further down the line, we will be surprised by what the use of AI in psychiatry can achieve, just as 20 or 30 years ago we would have been surprised if someone had claimed that smartphones were going to become such a big part of our lives, or that AI was going to become so prominent in academic discussion. And yet, we spend several hours a day on our smartphones and the potential benefits and downside of AI have become one of the most debated issues both inside and outside of academia.
