When you’re scheduled for a medical check-up, you might hear from Ana—a friendly voice designed to guide you through your appointment and address any urgent queries. Ana, with her soothing and approachable manner, is adept at comforting patients, similar to many nurses across the nation. However, Ana is unique in being accessible round-the-clock and conversing in various languages, such as Hindi and Haitian Creole. This is because Ana is not a person but an artificial intelligence (AI) program developed by Hippocratic AI, one of the several emerging enterprises that provide solutions to automate tasks traditionally performed by nurses and medical assistants.
This innovation represents a significant advance of AI in healthcare, where numerous hospitals now employ complex computer systems to oversee patients’ vital signs, detect emergencies, and execute care protocols—roles that were once managed by nurses and other medical professionals. Healthcare providers assert that AI aids nurses in operating more efficiently and tackling issues of burnout and understaffing. Nonetheless, nursing unions are concerned that this relatively misunderstood technology is overriding nurses’ knowledge and diminishing patient care quality.
Michelle Mahon of National Nurses United remarked, “Hospitals have waited for a tool that appears credible enough to substitute for nurses. The goal seems to be to automate, de-skill, and ultimately replace healthcare providers.” Her organization, the most extensive nursing union in the U.S., has orchestrated over 20 protests nationwide to advocate for a role in determining AI’s application and safeguarding against penalties for ignoring automated advice. Concerns heightened in January when Robert F. Kennedy Jr. proposed that AI nurses, as competent as doctors, could serve rural areas. Furthermore, Dr. Mehmet Oz, the nominee to lead Medicare and Medicaid, expressed that AI might emancipate healthcare staff from overwhelming paperwork.
Hippocratic AI had initially advertised a $9 per hour rate for its AI assistants, substantially lower than a registered nurse’s $40 per hour rate. The company has since withdrawn such comparisons, focusing instead on its robust service offerings and the thorough testing of its systems, though it did not provide a statement following an interview request. Despite the potential, AI in healthcare settings can occasionally produce false warnings and risky recommendations.
Hospitals have been engaging with technology designed to optimize care and reduce costs for years, incorporating sensors and motion-sensitive cameras. These tools, now integrated with electronic health records and scrutinized to foresee medical issues and direct nursing care, sometimes act prematurely, even before a patient is assessed by medical staff. Adam Hart, an emergency room nurse in Henderson, Nevada, experienced this firsthand when a computer system incorrectly flagged a sepsis risk in a dialysis patient, suggesting potentially harmful treatment. After challenging the directive and involving a physician, appropriate care was then administered. Hart noted, “Nurses must continue using their judgment; delegating our decision-making processes to machines is dangerous.”
Although nurses acknowledge AI’s goal to simplify the monitoring of multiple patients and expedite response times, the reality is often an influx of false alarms, sometimes mistakenly treating normal bodily functions as emergencies. Melissa Beebe, a cancer nurse, shared her struggles with distinguishing between meaningful alerts and false signals in light of the system’s frequent inaccuracies.
Understanding AI’s role in hospitals requires recognizing that even sophisticated tools cannot replace certain observational abilities, like detecting subtle facial expressions, as pointed out by Michelle Collins, Loyola University’s College of Nursing dean. Nonetheless, as estimated by the U.S. government, more than 190,000 nursing positions will be available annually until 2032 due to an aging population and retiring nurses, indicating the rising necessity of AI to support rather than replace healthcare professionals.
At the University of Arkansas Medical Sciences, AI programs assist in making numerous calls to surgery patients, gathering essential data before procedures amid time constraints. Dr. Joseph Sanford, overseeing health IT, sought AI solutions to alleviate the burden on his staff without incurring overtime expenses. Employing Qventus, an AI assistant, the hospital manages preparing patients with AI, maintaining transparency about when patients are interacting with AI or humans.
While serving administrative purposes, other AI companies envision broader roles for their technologies. The company Xoltar, specializing in humanlike avatars, collaborates with Mayo Clinic to create AI assistants that help patients with chronic pain management and smoking cessation by engaging effectively during video calls. However, AI may be more suitable for patients proactive about their care rather than those with extensive healthcare needs.
Nursing experts, like Roschelle Fritz, emphasize considering AI tools’ appropriateness for the very sick who dominate U.S. healthcare service needs. Augmenting care with AI appears promising, but as Collins stated, it remains crucial to ensure that technology enhances rather than replaces the essential human touch in healthcare.
