Experts Sound Off on How AI Became a Bad Buzzword

Ryan Black
APRIL 11, 2018


Buzzwords were a major theme of our coverage at last month’s HIMSS meeting in Las Vegas, Nevada. In the course of interviewing numerous attendees—from tech company CEOs to hospital Chief Medical Officers and cybersecurity experts—Healthcare Analytics News™ asked each one what their least favorite healthtech buzzwords were.

No term—not even blockchain—was mentioned more often than artificial intelligence. Some were technological skeptics: Medicomp’s CMO Jay Anders, MD, is always outspoken on this topic. He doesn’t think patients will fully trust AI-based decision support tools.

“If you're a woman and now you've been diagnosed by the machine with breast cancer, are you going to allow your breasts to be operated on because the machine said so?” he asked. “It never will breach the trust of a clinician seeing a patient, building that bond and actually treating.”

Others were more focused on the general misuse of the term—which many think is abundant (and at HIMSS, they might have a case).

“Artificial intelligence, broadly defined, is overhyped,” Sansoro Health CMO David Levin, MD, said.

“We're definitely at a phase now where it's overused. While there's substance there and there's things we have to achieve with that, more precision's needed, because the term is starting to lose its meaning,” Rush University Chief Research Informatics Officer Bala Hota, MD, said to Healthcare Analytics News™.

Before engaging any of what Levin called “more fantastical” ideas about AI, a lot of experts thought that the concepts should be applied to more practical means. It was the lack of practicality, 2bPrecise CEO Assaf Halevy said, that had turned AI into a buzzword.

“The thing that you should focus more on is how can we use these tools to analyze data that has proliferated beyond the ability of any one individual to fully be able to go through? How can you get a tool to help analyze all that and summarize or condense down information that can then be utilized by the actual professional?” healthcare attorney Matthew R. Fisher asked.

Of course, there are practical applications: The FDA just approved an AI-based system for diabetic retinopathy detection.

TRANSCRIPT:

LEVIN: Artificial intelligence, broadly defined, is overhyped.

ANDERS: Everybody seems to believe, and they still believe that artificial intelligence is going to solve healthcare treatment problems

Halevy: People again use it as such a broad term, not the practical way to think of it—you know, what can you really do with it—today that's one of those [buzzwords].

Hota: we're definitely at a phase now where it's overused and it's while there's substance there and there's things we have to achieve with that more precision's needed because the term is starting to lose its meaning.

Levin: If you wanted to find that as things like predictive analytics or maybe Siri like voice-activated assistance yeah I think that maybe in the in the near term

Fisher: artificial intelligence doesn't necessarily exist I think at least in the form that might be popularly thought of when you hear that term it's really only as good as the information that's there

Anders: there's a trust issue and there's a data issue. If you're a woman and now you've been diagnosed by the machine with breast cancer
are you going to allow your breasts to be operated on because the machine said so? it never will breach the trust of a clinician seeing a patient building that bond and actually treating.

Levin: Some of these more fantastical ideas? I think someday [will work]. But I'm also not sure that's where our priority should be right now, there's a lot of other low-hanging fruit in terms of just simple process improvement leveraging the technology we have today.

Fisher: The thing that you should focus more on is how can we use these tools to analyze data that has proliferated beyond the ability of any one individual to fully be able to go through? How can you get a tool to help analyze all that and summarize or condense down information that can then be utilized by the actual professional?”

Related Coverage:
Blockchain: Hype is Here, Use Cases Aren't
An Investor and an Innovator: The Birth of Sansoro Health
Dissecting the HIMSS Blockchain Buzz

SHARE THIS SHARE THIS
1
Become a contributor