Staying Healthy
Will artificial intelligence replace doctors?
Ask the doctor
Q. Everyone's talking about artificial intelligence, and how it may replace people in various jobs. Will artificial intelligence replace my doctor?
A. Not in my lifetime, fortunately! And the good news is that artificial intelligence (AI) has the potential to improve your doctor's decisions, and to thereby improve your health — if we are careful about how it is developed and used.
AI is a mathematical process that tries to make sense out of massive amounts of information. So it requires two things: the ability to perform mathematical computations rapidly, and huge amounts of information stored in an electronic form — words, numbers, and pictures.
When computers and AI were first developed in the 1950s, some visionaries described how they could theoretically help improve decisions about diagnosis and treatment. But computers then were not nearly fast enough to do the computations required. Even more important, almost none of the information the computers would have to analyze was stored in electronic form. It was all on paper. Doctors' notes about a patient's symptoms and physical examination were written (not always legibly) on paper. Test results were written on paper and pasted in a patient's paper medical record. As computers got better, they started to relieve doctors and other health professionals from some tedious tasks like helping to analyze images — electrocardiograms (ECGs), blood samples, x-rays, and Pap smears.
Today, computers are literally millions of times more powerful than when they were first developed. More important, huge amounts of medical information now are in electronic form: medical records of millions of people, the results of medical research, and the growing knowledge about how the body works. That makes feasible the use of AI in medicine.
Already, computers and AI have made powerful medical research breakthroughs, like predicting the shape of most human proteins. In the future, I predict that computers and AI will listen to conversations between doctor and patient and then suggest tests or treatments the doctor should consider; highlight possible diagnoses based on a patient's symptoms, after comparing that patient's symptoms to those of millions of other people with various diseases; and draft a note for the medical record, so the doctor doesn't have to spend time typing at a computer keyboard — and can spend more time with the patient.
All of this will not happen immediately or without missteps: doctors and computer scientists will need to carefully evaluate and guide the development of new AI tools in medicine. If the suggestions AI provides to doctors prove to be inaccurate or incomplete, that "help" will be rejected. And if AI then does not get better, and fast, it will lose credibility. Powerful technologies can be powerful forces for good, and for mischief.
About the Author
Anthony L. Komaroff, MD, Editor in Chief, Harvard Health Letter; Editorial Advisory Board Member, Harvard Health Publishing
Disclaimer:
As a service to our readers, Harvard Health Publishing provides access to our library of archived content. Please note the date of last review or update on all articles.
No content on this site, regardless of date, should ever be used as a substitute for direct medical advice from your doctor or other qualified clinician.