Artificial intelligence is no longer a niche tool relegated to data science teams or academic labs — it’s becoming a foundational pillar of modern healthcare strategy. Stanford Health Care is at the forefront of this transition.
The health system rolled out an AI 101 course to ensure all IT employees had baseline training in AI. Then Stanford rolled the program out to the entire organization along with a curated toolset for anyone to experiment with AI.
“The goal of that is really not to think about a team that does AI, or a person that’s responsible for AI, but rather everything and everybody is responsible for it, and needs to have some baseline understanding of it,” said Michael Pfeffer, MD, chief information and digital officer at Stanford HealthCare and Stanford School of Medicine, during a session at the Becker’s 15th Annual Meeting. “This has been really successful as we think about this point where we are moving to where our entire project portfolio is going to be a minimum of 30% of the projects we do must involve AI. The thinking is, if we’re doing a project the old way, let’s not do it the old way anymore. Let’s do it the new way. That’s been the philosophy.”
Outside of the IT organization, Stanford’s dean launched a RAISE health initiative, which stands for “responsible AI for safe and equitable health” as a guideline for the enterprise initiatives focused on AI. The goal is for everyone to understand the possible power and uses cases for AI in both medicine and education.
“We don’t want to just put AI into things,” said Dr. Pfeffer. “Part of that training is really identifying the problem you’re trying to solve and asking the question: should it be solved with AI or not? A lot of the stuff doesn’t need to be solved with AI. Getting to 30% of projects may sound like a low number, but it’s 30% of the projects that actually have a solution that is the right solution with AI. It’s been more challenging than we thought it would be.”
Stanford has created a secure portal with around 15 large language models and opened it to all network members. Anyone at the system can login to the portal and experiment with the LLMs and leverage new ideas quickly.
“What we’ve seen from that has been extraordinary, from incredible engagement to creativity to just overall excitement about the platform,” said Dr. Pfeffer. “For me, that’s really been one of the most exciting things to see people engage with this new technology in ways that in the past have been difficult.”
There are several ongoing projects at various departments using the LLMs, and Dr. Pfeffer’s team has been able to anonymously track usage. When prompted, the ChatGPT tool can break down the clinical and administrative use cases inspiring team members across the system. Dr. Pfeffer said the patient experience department in particular has had good results; the team reported ChatGPT gave them “two FTEs worth of people” by creating beautiful response letters to patients.
