The College Fix

‘Any part of medicine can benefit from teaching’ about AI, researcher says

Artificial intelligence and machine learning have become integrated into our everyday lives, from social media to cars to medicine. Yet, the curriculum for how to make use of this technology in medical school is still quite limited.

Medical schools have “little to no AI/ML training…Most curricula incorporate basic statistical training, but that’s about it,” said Michiel Schinkel, a researcher of the Amsterdam University Medical Centers, in an email interview with The College Fix.

In support of Schinkel’s analysis, Drs. Ozan Karaca, S. Ayhan Çalışkan and Kadir Demir report in the journal BMC Medical Education, “current medical curricula are not completely responding to the needs of future doctors in terms of preparing them for the AI era.”

Schinkel is bullish on machine learning. “Any part of medicine can benefit from teaching AI/ML. Radiology and cardiology and intensive care have been frontiers in the field, probably because they capture large amounts of high-quality data by nature,” he said.

“However,” Schinkel speculated, “the real value of [machine learning] might be in those situations with noisy and infrequent data, where it is hard to find any patterns for physicians.”

In other words, the artificial intelligence might pick up on something that humans miss, and bring that to the medical professionals’ attention.

The benefits of AI

According to many estimates, the integration of artificial intelligence and machine learning into medicine could be beneficial on multiple fronts.

First and foremost is patient care.

The use of machine learning in medicine would significantly improve population health “via continuous monitoring and coaching and will ensure earlier diagnosis, tailored treatments, and more efficient follow-ups,” argue editors Adam Bohr and Kaveh Memarzadeh in an extract of their book “Artificial Intelligence in Healthcare.”

They argue that artificial intelligence will facilitate moving from a treatment-of-disease focused model to a more management-of-health focused model. Patients will not only be better cared for but will likely also endure fewer hospital visits, treatments, and medical emergencies.

Another relevant consideration is the reduction of cost.

“It is estimated that AI applications can cut annual US healthcare costs by USD 150 billion in 2026,” say Bohr and Memarzadeh

If that is true, then not knowing how to effectively utilize these artificial intelligence applications could become a major roadblock in offering the best and most cost effective standard of care possible.

Humans still in driver’s seat

One roadblock to better knowledge of artificial intelligence in medicine is the fear that machine intelligence will take over medicine and edge out medical professionals.

However, that is not what the research is currently showing us about how artificial intelligence is being used in medicine.

Korean researchers Seong Ho Park, Kyung-Hyun Do, Sungwon Kim, Joo Hyun Park, and Young-Suk Lim found that artificial intelligence “tools for medicine mostly play the role of a virtual assistant for physicians and healthcare systems,” making the provision of care faster and more efficient.

They argue that is how it should be, saying, “In applying AI technology to patients, medical professionals are not ones who are in the backseat but should be in the driver’s seat.”

MORE: Berkeley scientists developing artificial intelligence tool to combat ‘hate speech’ on social media

IMAGE: Nestor_Rizhniak_Shutterstock

Like The College Fix on Facebook / Follow us on Twitter