Why CIO Jeremy Huval has to think about generative AI

The healthcare industry may not be fully embracing generative AI, but it’s at least giving the tech one of those “we’ve met” nods.

According to the Deloitte Center for Health Solutions, and the 2024 Life Sciences and Health Care Generative AI Outlook Survey, which polled 60 healthcare C-suite executives who work at companies with annual revenues of $500 million or more: 75% of leading healthcare companies are already experimenting with generative AI or attempting to scale across the enterprise, and 82% currently have or plan to implement governance and oversight structure for the technology.

Early ideas include the potential to “ease documentation burdens, handle pre-op workflows, and simplify insurance claims,” as Deloitte mentioned in another study.

That kind of gen AI experimentation from healthcare professionals keeps Jeremy Huval, chief innovation officer at the certifications provider HITRUST, busy.

“We help customers communicate that they’re secure. They’ve got their privacy processes in place to stakeholders. And so in the innovation group, I have to think about gen AI…because those have very specific risks that are different than traditional IT,” Huval said.

Huval spoke with IT Brew about what the applications and risks he’s seen emerge with the emerging tech.

Responses have been edited for length and clarity.

What healthcare tasks are especially suited for this kind of application of inputting something and getting an output quickly?

If I were starting up a new medical practice, gen AI can help me set up my website; it can write marketing copy for me; it can help me design my logos; it can help me actually write my commercials.

What about in the clinical space?

If I’m a doctor, and I’m recording my notes and interactions, right now I may have to actually sit back and transcribe all that into actual data entry into the system. That could go away. That could no longer require the step of me stopping and typing into the computer. That whole thing could be from my voice to the electronic medical record (EMR). And all I have to do is eyeball it for accuracy.

Gen AI, I look at it as interfacing with voice and text, and potentially images in the future, in the medical context, just to replace so much that’s happening manually right now.

Are there healthcare tasks that are not especially suited for gen AI?

Right now, you still need a human in the loop. The generative AI is synthesizing; it’s reading and understanding a lot of really complex data. But to actually trust the output of a generative AI with very-high-risk decisions that can impact someone’s life, or someone’s quality of life? We’re not there yet.

What would give healthcare employees confidence to use large language models?

When cloud tech came out, and healthcare companies were considering using it, there was strong resistance to ever putting any protected health information in any cloud ever.

Why? “I don’t know if it’s secure enough, I don’t know if I can still comply with HIPAA and other regulations if I put that data in that cloud. So, I’m gonna wait for it to be figured out.” There are still healthcare systems that have not fully embraced the cloud.

What’s funny about the AI revolution is we’re not seeing that level of resistance in organizations generally, including healthcare. There’s actually very much an eagerness to embrace this technology, and that creates the risk of people moving too fast, and putting too much trust and too much data in these systems, without having proper assurances that the security is in there.