Promising AI applications in healthcare: Chat GPT-4, natural language processing

Editor's Note

Chat GPT-4, the latest version of an artificially intelligent chatbot, shows promising applications in healthcare, as reported by The New York Times on March 14.

With greater accuracy and descriptive abilities, the chatbot was recently tested by Anil Gehi, MD, associate professor of medicine and cardiologist at the University of North Carolina at Chapel Hill. Dr Gehi described patient history and recent complications using several advanced medical terms, and asked Chat GPT-4 how the patient should be treated.

“That is exactly how we treated the patient,” he said. In testing other scenarios, Chat GPT-4 provided similarly correct answers.

Another technology advancement, natural language processing (NLP), is already being used in healthcare “to reduce manual entry, improve coding, and enhance research efforts,” as reported in the 2022 November/December issue of OR Manager. Strong results in the use of NLP have helped predict or detect adverse outcomes, risk of readmission, and patient discharge. New applications are being tested to streamline processes, strengthen patient data and electronic health records, and address preauthorization for services.

Part 1 of the two-part NLP series explores some of the uses of NLP in surgery. Part 2, published in the January 2023 issue of OR Manager, discusses the role of NLP in outcomes and general clinical research, as well as challenges of the technology and factors OR leaders should consider when purchasing products that use NLP.

Healthcare providers must be careful; “No tool or instrument is perfect,” says John Fischer, MD, of the Hospital of the University of Pennsylvania in Philadelphia.

Similar challenges are found in both NLP and Chat GPT-4. NLP relies on artificial intelligence (AI) algorithms trained by humans, which could cause unintended biases that skew outcomes, among other issues. Chat GPT-4 still experiences “hallucinations,” meaning it does not know what is true or false.

As healthcare facilities and providers continue to adopt AI, it is important that a human makes the final judgement calls on patient care, the report says.