AI:‘The Future is Actually Already Here for Clinicians’

“Clinicians are already using AI; the problem is they don’t know they are using AI,” states Deepti Pandita, MD, FACP, FAMIA, chief medical information officer, University of California Irvine. “Your EMR may be presenting you with a lot of clinical support tools that behind the scenes are driven by AI. They have machine learning algorithms that are constantly working.”

Pandita presented a scientific session, Artificial Intelligence/Machine Learning: Practical Applications in Medicine, at this week’s ACP conference being held in San Diego.

With the enormous amount of responsibilities that providers have including remembering patient information, paperwork, increasing workloads at institutions, Pandita sees the potential to alleviate the need to remember everything. “AI helps with reducing the cognitive burden.”

The concern, Pandita says, is if the AI is set up to address individual patients. “The question is, we don’t know if whether those AI algorithms are configured in a manner that are truly safe for your patient population, and how do you build confidence in those?”

As with all evolving technologies it remains to be seen what the intended and unintended consequences will be, but Pandita is confident human physicians and other providers will not be replaced. “I think the human brain surpasses anything AI can do.”

She sees it as artificial augmentation and not artificial intelligence. “You are augmenting what the human brain can do and adding artificial intelligence to it…think of it as a helper in the clinical setting.”

She uses the example of how AI, specifically chatbots, can now help with writing notes and possibly interactions with patients in showing more empathy. This last concept was shown in a new study published in JAMA Internal Medicine, which demonstrated that chatbot responses were preferred over physicians’ responses in a social media forum where patients’ questions were posted.

Contagion spoke to Panditia about some of the ideas we can expect going forward with this emerging technology, such as writing scientific papers, and the need for trust-building with AI as it will only become increasingly omnipresent.