Chatbots in the healthcare sector are rapidly increasing in scope and popularity, and it’s easy to see why: they provide personalised responses to clinical questions which tend to be far more useful and accurate than a patient-led internet search. But is it safe to rely on this kind of artificial intelligence to deal with the health of a population?
There are some areas of concern regarding this type of artificial intelligence. Despite the use of natural language processing (NLP) and conversational interfaces (CI) to humanise the chatbot experience, many patients say they miss an empathetic and real connection with a professional. Some just don’t trust the chatbot to interpret their symptoms correctly using an algorithm, or are concerned that their sensitive medical data won’t be handled or stored securely. There’s also the possibility of a serious error due to the patient not attributing significance to certain symptoms, or them simply selecting the wrong option by mistake due to a lack of familiarity with the technology.
However, there are a range of benefits to them too — the vast majority of healthcare chatbots are available 24/7, have language options at the touch of a button, and are free to use. These features help improve access to medical advice for anyone with an electronic device (which is particularly relevant in countries where private medical insurance is required). If an automatic record entry system is enabled via the chatbot, this can reduce the amount of time spent documenting and inputting responses. They can remove any issues around patient’s need to travel, take time off of work, or find childcare in order to attend a face-to-face appointment. They can also provide assistance to those in rural communities who live far from a healthcare setting.
The use of chatbots is understandably inferior to the gut instinct of a healthcare professional when they know ‘something just isn’t right’. However, they are designed to complement the work of healthcare staff rather than replace them, and for that purpose their potential is far-reaching. They are quickly developing into an integral part of our medical landscape, and we should certainly be prepared for them to feature as a familiar tool in the virtual doctor’s bag.