Less than one year after ChatGPT's debut took the digital aspect of many industries by storm, 18 percent of healthcare employees say they use the tool frequently, according to a July LinkedIn poll conducted by Becker's.
Of the 1,617 respondents, 23 percent reported they use the tool but do so infrequently, and 58 percent reported not using it at all. Becker's does not have detailed insight into the respondents' roles or organizations in LinkedIn polls outside of the information that is publicly available.
"You can use it to answer a lot of questions. You should not use it to DO your work," Monica Schmidt, PhD, executive director of health economics and health equity analytics at Cone Health in Greensboro, N.C., said in the poll's comment section. "This is a powerful, shared decision-making tool. It outperforms a Google search when you have a quick question. Always good to check references and confirm accuracy."
Kristen Rasmussen, MSN, RN, senior innovation catalyst at Lee Health in Fort Myers, Fla., commented that she uses the tool "every day" and that it has allowed her to be "twice as productive, and [I] can use my gained time for activities that really benefit from my expertise."
While the tool may be useful, experts caution healthcare professionals about what information they choose to put into ChatGPT and to treat the tool like a business associate and approach it with security practices in mind.
"Generative AI tools like ChatGPT need to be treated as a business associate under the HIPAA security rule because some might not realize the risks of putting patient data into it," Mark Bowling, a former chief information and security officer for hospital systems and the current chief information and risk officer at cybersecurity company ExtraHop, told Becker's. "These tools should be treated exactly the way you would treat an outsourced EMR such as Allscripts or eClinicalWorks."