One wonders what John McCarthy had in mind when he introduced the term Artificial Intelligence in 1956. I do not like the term Artificial, and prefer Augmenting, Amplifying, Accelerating and Assisting.
AI is only an extension of natural intelligence. It’s the use of computer systems to perform tasks normally associated with human competencies. These include pattern and speech recognition and visual perception leading to decision making. COVID-19 has resulted in “contactless” healthcare, with clinical distancing becoming the new norm. This is leading to a global behavioural modification and technology acceptance in healthcare, changing our mindset to believe things can be equally good, if not better, in remote mode.
AI is the new electricity, the buzzword everywhere, and will now radically and exponentially transform healthcare.
A disruptive technology, AI will have an impact in healthcare that cannot be underestimated. Using it in public health and to manage patients is no longer a question of if — it is a question of how and it is a matter of now! Machine learning, the basis of AI, gives computers the ability to learn, unlearn and relearn without being explicitly programmed. From simple pattern recognition, today’s algorithms learn unsupervised, from unstructured or unlabelled data. Predictive analytics and deep mining can be used in hundreds of clinical situations, be it in diagnosis or recommending optimum management. Fuzzy logic, the basis of all clinical decision making, is similar to human thinking and interpretation.
In the real world, clinical scenarios are not just “present or absent”. Symptoms are described as “never, rarely, sometimes, often, most of the times, always, perhaps” and so on. Each specific symptom can be “mild, moderate, or severe”.
Fuzzy logic can quantify these multiple variables giving appropriate weightage. Tomorrow’s 5P (predictive, personalised, precision, participatory and preventive) medicine will have AI as a major component. In the post- COVID era, AI will cause a paradigm shift in healthcare delivery. It can unlock information hidden in massive amounts of data. However, using AI for a significant clinical impact for an individual patient, under specific socio-economic conditions factoring patient desires, will be a challenge. AI requires thorough and systematic evaluation prior to integration in clinical care.
Uncertainty principle
The good AI-enabled physician should never forget that medicine is still a “science” of uncertainty. For centuries, the essence of practising medicine has been obtaining as much data about the patient’s health or disease as possible and taking decisions. “Listen, listen, listen, the patient is telling you the diagnosis,” Sir William Osler said. I wonder whether the body language of a humanised chatbot will replicate that of this great physician. Wisdom then presupposed experience, judgment and problem-solving skills using rudimentary tools and limited resources.
Fifty years ago, Lars Leksell, the inventor of the Gamma Knife, said, “A fool with a tool, is still a fool.”. When one has a hammer everything around you looks like a nail, particularly if the hammer is expensive! As one trained by the ear before computers entered the mainstream, I am concerned that AI may not just be a means to achieve an end. It could become an end by itself. Time alone will tell if AI in healthcare will be a bane or a boon.
To the clinician, notwithstanding peer pressure, AI should become an integral part of his armamentarium, when there is evidence that outcomes are better with reduced costs. We are in a stage of transition. All transitions offer great opportunities. AI should never dehumanise medicine, never replace a commiserating clinician. It’s hoped that AI-enabled clinicians will spend more time practising TLC (tender loving care), empathising and sympathising with those who need it the most, rather than getting drowned in voluminous data and paper work.
(The author is a past president of the Neurological Society of India and the Telemedicine Society of India )
drganapathy@apollohospitals.com