Plenty of buzz for AI in healthcare, but are any systems actually using it?

It doesn’t take much to grab hospital executives’ attention these days—just casually mention the words “artificial intelligence” and watch their eyes widen at the thought of the technology’s untapped possibilities.

The promise of AI has hit the healthcare industry full force, in part because industry leaders and researchers are searching for ways to utilize a new ever-growing database of patient information.

In some ways, the promise is very real. Researchers have shown that AI can be particularly useful in augmenting image-based specialties like radiology and pathology, and machine learning has been shown to detect skin cancer with the same accuracy as a board-certified dermatologist.

RELATED: AI could prompt merge of radiology, pathology into one specialty

Hospital CEOs and venture capitalists at some of the country’s foremost medical systems have taken note. Thomas Sudow, director of business development at Cleveland Clinic Innovations, the system’s venture arm, told MedCity News that he sees AI as a compelling new technology, particularly as a clinical decision support tool. Kaiser CIO Dick Daniels told CIO that AI is “the one technology that could make a huge and dramatic change" because it can digest large amounts of data quickly, and turn that into usable information for clinical decision-making.

Mayo Clinic CIO Christopher Ross told MedCity that the system has “a number of projects underway” including a partnership with IBM Watson, but cautioned that the technology is still in a toddler phase because it relies on “supervised learning.”

RELATED: Increasingly powerful AI systems are accompanied by an 'unanswerable' question

Not everyone shares the same regard for systems like IBM Watson. Chamath Palihapitiya, CEO of the venture capital firm Social Capital, created some buzz this week when he called Watson “a joke,” adding that IBM uses its powerful “sales and marketing infrastructure” to get companies with “asymmetrically less knowledge to pay for something.”

Palihapitiya later told CNBC he should have been “more careful with my words,” but said he would like to see his company’s AI technology go up against Watson “head-to-head.”

For all the possibilities that healthcare leaders see in AI, there are few practical applications currently in use because the technology hasn’t been aligned with the workflow and priorities of clinicians. Data scientists at Booze Allen Hamilton and an emergency physician at Georgetown University School of Medicine highlighted these shortcomings in an article for Health Affairs, highlighting the fact that EHRs don’t have the capability to use machine learning or cognitive computing, and a lack of robust outcome data is holding back third-party applications.

In the short term, the most practical applications for AI may reside in taking on “mundane” data-driven tasks, like reviewing patient charts. To make use of AI’s capabilities in the long term, the healthcare industry needs to find ways to “feed extremely data hungry” with more robust and useful information on clinical outcomes that aren't typically found in the EHR.

“While we believe machine learning holds great promise, it is far from clear how it will transform health and health care in the short to mid-term,” the researchers wrote. “Today, policy makers and industry executives face decisions about when and how to invest in machine learning to optimize organizational effectiveness and efficiency without wasting capital funds on premature or nonvalue-adding technologies.”