Artificial intelligence algorithms touted as having the ability to outperform human clinicians in disease prediction and diagnosis may ultimately hurt, rather than help, healthcare, according to Enid Montague, PhD.
In an op-ed for The Hill, Dr. Montague, an associate professor of computing at DePaul University and adjunct associate professor of general internal medicine at Northwestern University, both in Chicago, described the dangers of introducing free-standing AI into healthcare.
While any developments that reduce clinicians' administrative burdens and improve diagnoses and outcomes are certainly a positive, according to Dr. Montague, any AI introduced in a healthcare setting should be done so to augment clinicians' expertise, rather than completely replacing it.
"Replacing humans with technology inappropriately could lead to complacency from physicians and reduced engagement from patients," she wrote, explaining that algorithms can contribute to increases in medical error, eliminate much-needed human empathy from healthcare and perpetuate racist and sexist biases.
"Yes, well-designed AI in healthcare systems can transform the health and well-being of members of society by allowing healthcare professionals to provide better quality care to more people and restoring balance to the people who dedicate their lives to providing care," Dr. Montague concluded. "But the danger is that AI systems that pit humans against algorithms will likely introduce new biases and errors into the U.S. healthcare system that will not only exacerbate health disparities but also make healthcare more dangerous for everyone."
More articles on AI:
Bon Secours Mercy Health invests in behavior change AI platform
Top 10 AI applications for healthcare in 2020: Accenture report
Physicians see 25% of their work being automated in the future, survey finds