Part 5: Artificial Intelligence is here to help in healthcare but it might take some time

Artificial intelligence has been utilized more and more in recent years, in nearly all aspects of life.

And with the development of ChatGPT and other AI software, medical industry leaders and experts are quickly discovering the answer to the question: "When will AI enter the healthcare field?"

This was a topic the Texas Tech Health Sciences Center explored at its Telehealth Week Conference last week in Lubbock, with a panel called "Opportunities for AI in Healthcare and Related Challenges."

Before talking about AI entering the healthcare field, people need to understand the basics of AI.

What is AI?

IBM defines AI using the mathematician Alan Turning's definition of "a system that acts like a human."

In the simplest form, AI combines computer science and data to have a program solve a problem in a way that thinks like a human.

Richard Greenhill, director of the Bachelor of Healthcare Management program at TTUHSC, said it is easier to think of AI in two categories.

"Generative AI is what you see all the fuss about now with ChatGPT and other things out there — large language models," Greenhill said. "Traditional AI is for those of us that are data scientists and work in that area where we actually ask the algorithm to do something, look at some data."

When talking about AI, one might hear the term "Big Data" — which refers to the vast amount of data generated and produced by the everyday person.

So how will AI and Big Data impact healthcare?

There are no clear answers. The panelist spoke about leveraging AI to create higher-resolution CT scans or help diagnose patients, but there are a few hurdles in the way.

One hurdle is the fact that right now, there isn't much data to generate an AI program from, said Courtney Queen, assistant professor at TTUHSC.

"We actually don't have enough data to do the things that we need to do already," she said. "Thinking about technology development and screening and diagnostics, we just can't get a hold of the data that we need."

The panelist also said another complication with using big data is that someone has to clean, or process, it — especially if someone inputs incorrect data.

Alan Pang, a TTUHSC surgical resident, used the example of misdiagnosing a patient and putting it on their record, and the implications that could have on the patient's future care.

"You'd be surprised how hard it is to get that initial wrong diagnosis out — it's there forever," he said. "You'd have to go into every chart or develop some way to go into every chart and say that is a wrong diagnosis and I'm going to ignore it because no matter what (electronic medical record system) that we're working with, it's just impossible to get that old diagnosis out of there."

The panelists agreed that we might see smaller versions of AI enter healthcare — like an at-home diagnostic system — but there will always be a need for patients to have some level of interaction with human medical experts.

This article originally appeared on Lubbock Avalanche-Journal: AI can help in healthcare, but it might take sometime