How to Integrate Chatbots and other Conversational Agents into Health Care

The technology for developing artificial intelligence-based chatbots (or, more formally, “conversational agents”) has existed for some time, yet its widespread implementation in clinical settings has so far been hampered due to logistical issues and ethical concerns.

Now, a new viewpoint piece out in the Journal of the American Medical Association proposes a framework for deploying this technology in a way that maximises social and health care-related benefits, and minimises harm.

If implemented effectively, chatbots could lead to improved health and reduced health care costs. Image: John Jackson via pexels.com

The authors argue that initiatives to use chatbots in hospitals and other health care facilities must first recognise that much of the data on their usefulness comes from research, not clinical implementation. “Knowing that, evaluation of these systems must be robust when they enter clinical space, and those operating them should be nimble enough to adapt quickly to feedback”.

In the paper, the authors lay out a list of considerations to be taken into account whenever a health care facility decides to make use of chatbots. These considerations include questions such as, “Who should monitor the chatbots and how often?”, “Are patients likely to take responses from chatbots seriously?”, “Who should be accountable for chatbot failures?”, “Is the task better suited for a chatbot or a human professional?”, and many others.

“To what extent should chatbots be extending the capabilities of clinicians, which we’d call augmented intelligence, or replacing them through totally artificial intelligence?” said co-author Ross Koppel. “Likewise, we need to determine the limits of chatbot authority to perform in different clinical scenarios, such as when a patient indicates that they have a cough, should the chatbot only respond by letting a nurse know or digging in further: ‘Can you tell me more about your cough?’”

Chatbots that are based on natural language processing algorithms present the opportunity to both improve health outcomes and reduce the cost of health care, yet their effectiveness can be ensured only given continuous evaluation and research.

Source