John Brownstein, PhD, chief innovation officer of Boston Children's, told Becker's that the organization is hiring a prompt engineer to work on large language models such as ChatGPT so that it can identify use cases and train its workforce on the most appropriate uses for the emerging technology.
"Artificial intelligence-based tools have the potential to increase efficiency and reduce burnout, and we believe one of the best ways to use these tools is to hire people who have expertise in the field," Dr. Brownstein told Becker's. "This role can help Boston Children's build and engage with these models."
The engineer will be part of the hospital's Innovation and Digital Health Accelerator team.
The person will design and develop prompts to effectively gather data from generative AI programs and refine the models for healthcare-specific applications.
Dr. Brownstein said by having an in-house expert, Boston Children's will be able to begin building out some of these tools internally, identify use cases for the technology and train its workforce on the most appropriate use of these tools.
But when it comes to generative AI programs, fewer than half of healthcare organizations are planning to pilot a program around the technology in 2023. This is because hospitals and health systems have to be careful when it comes to how they are using it.
"When it comes to reducing tasks such as clinical documentation or analyzing data sets, AI applications work great," Dr. Brownstein said. "But, when you're asking AI to generate de novo insights — whether that's clinical or research — you have to check its sources."
Dr. Brownstein said Boston Children's is advocating for use cases of generative AI in which the technology can help do the activities that can lead to physician burnout, such as navigating large quantities of data, summarizing information and generating text out of "noisy data."
"We haven't seen this kind of level of innovation since the search engine or iPhone," Dr. Brownstein said. "For instance, using AI to respond to patients' emails or to generate post-discharge notes automatically will help bring back a love of practicing medicine for many of our providers."