Employees of Harris Health System in Texas are allowed to use ChatGPT or Grammarly—two forms of artificial intelligence (AI)—but their use is monitored by the information security team, and any red flags, such as the inclusion of patient information, are escalated to the compliance team to investigate. It’s the price of access to the AI tools, which almost didn’t happen.
“We had long conversations about whether we should block them,” said Carolynn Jones, chief compliance and risk officer at Harris Health. Ultimately, it decided some uses of generative AI have value “but we needed to establish guardrails,” she said. “We are trying to have a more formalized governance process” for approving AI and spreading the message to the workforce about how to use it appropriately.
To move its AI framework forward, Harris Health has established a governance committee that includes Jones, the chief information officer, chief technology officer, chief information security officer, chief legal officer, chief medical officer and chief nursing officer. “It’s meant to be a strategic group to set the organization’s position and policy on AI,” she explained. “We have to catalogue any systems that are utilizing some version of AI and determine what our requirements are for the use of AI.”
At the same time, Harris Health is considering the effect of AI on morale. For example, with AI technologies that provide preliminary reads of imaging and chatbots that help in call centers, employees may wonder if their jobs are on the line, Jones said. “What are the possibilities for AI and what does it mean for individuals doing their jobs? We have to address that from a people perspective,” she said. “It’s really about trying to find how to most effectively and efficiently use the skills of the workforce.” It’s helpful to focus on “using AI as a way to complete tasks so individuals can perform other tasks at the top of their licenses.”