Can AI Save Modern Medicine? Or Will It Be the Other Way Around?

Medicine and artificial intelligence (AI) are very similar fields in some aspects: they both seek to help and aid humanity, improve day-to-day conditions for people, and continue the evolvement of their respective fields for the sake of their users.

So it’s no surprise that when we combine these two fields, the result is something grander than we’ve previously known. When AI is implemented in medicine, doctors and patients spend more quality time together, diagnoses improve, and treatments end more positively. When medicine’s principles are implemented in AI, AI becomes more human, just, and empathetic.

The Quick Doctor’s Appointment

Abraham Verghese is a physician and author; he spoke at the 2019 Singularity University Exponential Medicine conference in San Diego about how AI applications can improve modern medicine and allow doctors to get back to caring for their patients.

For most of us, visiting the doctor is a one- to two-hour activity that entails spending only 10 to 20 minutes with the actual doctor. You may have more questions, but you feel burdensome and reserve the follow-up questions for a Google search later that night. The doctor rushes back to his or her office and updates your electronic medical record (EMR), a task that takes as long as your time spent with them in the exam room. This pattern repeats all day long for the physician.

The huge emotional disconnect between the doctor and the patient in this interaction leaves both parties feeling incomplete. It can result in integrity loss: the doctor isn’t fully present mentally, and as a result, compliance or concerns from the patient can get lost in the shuffle.

Verghese says that doctors spend an hour updating EMRs for every hour they spend with patients, citing that EMRs cause most physicians more distress “than any other invention that’s happened in modern medicine.”

Verghese notes that providers may even feel so deflated that they consider themselves demoted to the position of a data clerk.

The Physician’s Lack of Support

Although many doctors feel animosity towards their particular EMR system, it’s nearly impossible for them to communicate this to each other in a suitable way. There is no mass-communication method for doctors to voice their concerns and discuss possible solutions.

Most patients think that the American Medical Association (AMA) is the voice of doctors, but AMA enrollment is actually not mandatory for physicians. At the end of 2018, there were 1,341,682 physicians, residents, medical students, fellows practicing in the U.S. But only 250,253 were AMA members. Almost 50% of the 250,253 members were students and residents.

Eric Topol is a physician who manages Medscape, hosts the Medicine and the Machine podcast, and founded the Scripps Research Translational Institute. Topol is a major advocate of a doctor’s union, one that harkens back to the olden days where unions were used as a vessel for fighting for rights and the contentness of its members. Medicine has become a multi-billion-dollar profit-driven business, where the patient and provider are the last priority.

Technology as a Partner

Verghese notes that physicians haven’t been able to control their work style due to a loss of autonomy. According to Topol, a doctor’s union would help drive better technology for physicians, giving them back satisfaction and happiness in their jobs. As a result, the quality of care for patients would grow.

Instead of framing AI as the technology that will take away doctor’s jobs (even though we’re experiencing a chronic shortage of medical professionals), we should view AI as a partner to healthcare.

To get back to the days where a doctor would spend more time learning about their patients and their family history, AI would be an indispensable technology. It can identify images faster than humans, often at equal or better accuracy; it can point out abnormalities in tissue slides; it can even help providers diagnose illnesses in real-time in some cases. Ultimately, AI affords us something very valuable, that physicians struggle to grasp onto: the gift of time.

Medicine’s Gift to Technology

As AI continues to find itself in more and more medical applications, we must not overlook the fact that AI ethics hasn’t grown at the same rate. Ethics guidelines and standards in AI are crucial to ensure there is no abuse of the technology for greed, profit, or personal gain.

Although AI has no equivalent in medicine for patients, it doesn’t mean that the principles of medicine can’t inform the ethics of AI. The true challenge will come from enforcing the ethics, figuring out who exactly would need to be prosecuted (i.e. would it be the developer team, the company who employs them, or the client they worked for?), and adjusting these guidelines for each subset of AI.

It’s true that, due to the diversity of AI developers, their backgrounds, their upbringings, and their locations, it would be nearly impossible to have everyone agree to a set group of ethics guidelines. Yes, each developer would have their own interpretation of the regulations and different ways of implementing them.

But if there’s even a slight chance that it could help protect consumers and end-users from abuse, bias, and harm, then we should try our hardest to publish an ever-evolving group of standards.

AI and Modern Medicine Depend on Each Other to Thrive

Because of Topol’s work, a new alliance named Osler’s Alliance has formed. It works to protect the doctor-patient relationship, focusing on what doctors can do today to improve care for tomorrow.

And medicine’s tradition of applying ethics to every level of the patient experience is a great template for AI developers and companies to follow in their development of new AI applications.

It turns out that without AI, medicine will grow slowly, and without medicine, AI will be a lot less human and just.