Artificial intelligence in all of its forms--not just generative AI--is going to transform healthcare, but the sector needs to be judicious about how it integrates AI, machine learning, deep learning and generative AI to lower costs and improve patient outcomes.

That broad takeaway was delivered by Anand Iyer, Chief AI Officer at Welldoc, in a DisrupTV interview. Iyer will be a speaker at Constellation Research's AI Forum Sept. 23 in New York.

Here are the takeaways from the Iyer chat with Constellation Research CEO Ray Wang.

Healthcare's intersection with technology and AI. Iyer said rising costs for healthcare and worsening outcomes are driving the need for innovation and technologies that can improve care and lower costs. Iyer said:

"Healthcare is becoming very expensive at the individual level, at the corporate level, and certainly at the macroeconomic level. It's such a large portion of the entire country's GDP. At a high level, healthcare is looking to do two simple things. Innovate. And move the continuum to focus efforts on management and prevention. Can I use data insights? Can I use modeling? Can I use the vast data from sensors to focus the attention on prevention? There are also inefficiencies throughout the healthcare systems. I think AI has the ability to actually have your cake and eat it too."

The noise around AI and machine learning vs. generative AI. Iyer explained:

"There's so much excitement, which is a very positive thing. But with excitement, of course, comes the misunderstanding of AI. AI is oftentimes synonymous with generative AI. AI is actually a much bigger thing. And it starts with what we call the rules-based engine. So, teach a machine to do what a human would otherwise do, and that can be accomplished with a simple set of rules or complex set of rules. There's even the gradation of the complexity of a rules engine. But the pluses of those types of AI approaches, of course, is that they're traceable. For a given rule, the output is known. And by the way, the regulatory agencies love that since they can trace a patient safety all the way back to a set of rules.

The next level of AI is using machine learning so is not fixed anymore. You're feeding data to an algorithm and it is adapting and changing. It's optimizing itself for different conditions, situations and patients. After that it is deep learning and that works well on large data sets, neural network models and then generative AI where it comes back with new content.

I think the reality is all four of these forms of AI, not just generative AI work their way into healthcare."

AI risk assessment in healthcare. Iyer said healthcare organizations need to integrate AI based on risk and intent. There's efficiency and then there's outcomes and risk. There are low risk categories like prevention and education and high-risk areas like mediation dosing. "You have to make sure you have the controls, the safety and the traceability to evidence-based guidelines," said Iyer.

Will AI replace a doctor? Iyer said:

"People ask us all the time, will AI replace a doctor? Hell no, absolutely not. But a doctor who uses AI may very well replace one who doesn't. The analogy that comes to my mind is when you look up in the nighttime sky and you see a point of light, you look up and say, 'oh, it's a star.' The more astute will say it's a planet. But then you look at that same point of light through the Webb telescope, and you see seven galaxies. The Webb telescope is to the astronomer, what AI and digital health is to the doctor. It lets the doctor see things that can't be seen by the naked eye alone. It lets them look deeper than what they can see in a five- or 10-minute office visit."

"There's no blame here. Especially primary care that's the front line for everything. Doctors are being bombarded and in many cases the evidence-based guidelines are changing rapidly and you can't keep up. Patients want the guidance so they don't want AI to be the doctor. But if I could have something that helps my doctor understand what's actually happening with me personally then AI is a transformation agent."

Simply put, Iyer is betting that AI can be like turn-by-turn instructions on Google Maps for physicians.

Personal electronic health records and data sharing. Today, patient records are locked into systems. "At the end of the day if I'm generating all this data as the patient I own the data," said Iyer, who noted that patients will want precise guidance, but also contribute to a larger model.

"I think the future is going to be one where there's going to be more collaboration with sharing data--anonymized of course and protected," he said. "Within those constraints, there's so much we can do by merging data sets."

 

Business Research Themes

Data to Decisions, Next-Generation Customer Experience

Roles

Chief Analytics Officer, Chief Data Officer, Chief Information Officer, Chief Technology Officer

Tags

#healthcare #cio, #AI #CDO #CIO #Analytics #CX, #GenerativeAI