New York - Experts at the HIMSS AI in Healthcare Forum said AI is a great tool to support clinicians, but it should never replace the human element of care.
As a patient, said Antonio Nunes, senior manager at AMD, “I don’t want to talk to Dr. AI.”
“It will never take over compassion,” said John Doulis, vice president of Data Services & Technology Innovation at HCA Healthcare.
The headline of “AI versus doctors” is the wrong way to think about it, said Ethan Goh, executive director at Stanford ARISE (AI Research and Science Evaluation) Network, Stanford. “Because everyone knows that's not true.”
The panelists spoke Friday at the HIMSS AI in Healthcare Forum, “From Pressure to Practicality: AI Use Cases Supporting Nurses and Physicians.” Justin Collier, chief technology officer at Lenovo Healthcare, moderated.
AI is a useful tool, the panelists agreed. But AI has revved up from AI 1.0 to 3.0 so quickly that no one wants to be left behind. Its inappropriate use can create trust issues.
AI in part has become an arms race, Goh said.
“There's a lot of technology that's adopted for the healthcare system, which may or may not work, which loses trust,” Doulis said. “The two physicians here spoke about that in hallucinations. So there has to be adoption in smaller chunks. Now you don't have the choice not to invest in AI. This is going to happen because AI equals automation. Automation equals lower cost. It has to be done.”
ADOPTION
Trust has to prove itself by not impacting the nursing workflow or the physician, Doulis said. When it’s not disrupting the workflow, that's where you see the adoption, he said.
Janna Templin, chief AI technologist at Lenovo, practices part time in criminal care ICU, cardiac, neuroport medicine and clinician technology.
“My peers, they don't ask for more data, they don't ask for more tools,” Templin said. “And right now, AI has a wonderful application of tools.”
When AI solves a problem that’s actionable, that’s been a personal revelation, she said.
Goh said AI can take on chart review and prior authorization. “So to me, at least, I think that's a really exciting frontier.”
Taking out the variability in care, standardizing care has been a way to save patients lives. Is the next frontier, they asked, an AI platform to predict sepsis?
Agentic AI is the hot topic, Doulis said.
“This will carry itself eventually through the clinical side,” he said.
“Technology is something that's been a constant the whole time,” Nunes said “We throw technology at everything. OK, despite shortfalls and budgets and hospitals, there's always money for investment in technology. … As trust manifests itself in investment, it will carry through.”
Email the writer: SMorse@himss.org