Healthcare & Technology

Why AI in Medical Diagnostics Is Both a Promise and a Warning

By Lesandu Wickramasinghe, YouthCare Connect

When you consider consulting a doctor, it typically implies conversing with the actual person who hears your issues and provides you with thoughtful responses. But what if a machine takes your frantic call instead of a human? A machine may contain the knowledge of thousands of people, but it does not have the empathy of a single human being.

Though modern technology is resourceful and efficient, we should not allow ourselves to forget about the human side of medicine that a machine could never replicate.

A Technology That Is Already Here

AI in healthcare is not a distant idea. It is already used globally, at a scale most people do not realize. Doctors currently consult with AI to review cases or keep track of patient records. Furthermore, the US Food and Drug Administration provided consent for 950 pieces of AI medical equipment — including AI and machine learning in radiology — back in 2024. Additionally, studies show that doctors’ use of AI has increased significantly in just two years, with percentages rising from 38% in 2023 to 66% in 2024.

Though these numbers are high, they do not explain whether this transformation is beneficial or harmful for patients.

What AI Gets Right

AI has one area where it excels: recognizing patterns in large sets of data faster than any human being can. In radiology, for instance, AI tools can rapidly and accurately identify early signs of diseases such as lung cancer, stroke, and breast cancer. Johns Hopkins Hospital, working with Microsoft Azure AI, trained algorithms to predict patient outcomes — including disease progression and readmission risks — by analyzing electronic health records and genomic data.

This matters because early detection is a crucial factor in efficient healthcare. A missed diagnosis today can become a crisis tomorrow. If AI can catch what a tired doctor working a 12-hour shift might overlook, that is not a small thing.

Where to Get Worried

The concerns around AI in medicine are real and documented:

AI systems learn from historical data, and historical medical data is tied to inequalities. If an algorithm is trained mostly on data from certain populations and diagnoses, it may not perform well across the full spectrum of patients. The communities already underserved by healthcare are the same ones most at risk of being failed by a biased algorithm.

YouthCare Connect’s Perspective

We are not saying that AI should be removed from medicine altogether. We are saying that its role should be an appropriate one. In our view, the best scenario is one where AI assists doctors rather than making life-altering decisions on its own. Radiologists are already getting help from AI when reviewing medical images. ER doctors use AI to identify which patients need the most immediate attention. AI tools also help surgeons with operative accuracy.

Medicine is not a software program you can simply download and run. If a computer application crashes, you may lose your files. If a diagnostic AI fails, a patient may lose their life. The algorithm is already in the exam room. The question is whether we are watching it closely enough.