OpenAI launched ChatGPT for Healthcare on Jan. 8, marking a significant step into the healthcare industry as some of the nation’s largest health systems begin piloting and committing to the platform.
ChatGPT for Healthcare is OpenAI’s enterprise platform designed for healthcare organizations, allowing clinicians, administrators and researchers to use generative AI tools within a centralized, secure workspace. OpenAI said the platform includes role-based access controls, governance features and single sign-on support. It runs on GPT-5.2 models evaluated through physician-led testing and real-world clinical scenarios.
Early adopters include Boston Children’s Hospital, AdventHealth (Altamonte Springs, Fla.), Baylor Scott & White Health (Dallas) and Stanford Medicine Children’s Health in Palo Alto, Calif. Many of these organizations had already worked with OpenAI on earlier AI capabilities, making the move to an enterprise-level ChatGPT platform a natural next step.
“We’ve been working with OpenAI for several years now,” John Brownstein, PhD, chief innovation officer at Boston Children’s, told Becker’s. “Internally, we launched a customized ChatGPT tool at the hospital that’s now used by more than 30% of our employees, and we’re seeing significant enthusiasm around these tools.”
Boston Children’s has also collaborated with OpenAI on research initiatives including genomics and rare disease diagnostics. These research efforts, according to Dr. Brownstein, are part of what makes this “a natural evolution” of its existing partnership. With ChatGPT for Healthcare, Boston Children’s aims to support clinician decision-making, reduce manual administrative work and help researchers with literature reviews and data analysis. The hospital is measuring success by utilization and return on investment.
“What tasks can be shifted to AI, how we can reduce costs for the health system, and how we can help ease burnout by taking work off people’s plates,” he said. “We have a dedicated analytics team focused on measuring that return, and much of our current reporting centers on hours saved across the organization. As we roll out this tool, we’ll be closely tracking whether and how we’re able to move the needle on those metrics.”
At AdventHealth, the path to ChatGPT for Healthcare began with a pilot in May using OpenAI’s ChatGPT for Enterprise, the business-facing version of the tool.
“We wanted to understand whether there were meaningful business use cases inside a healthcare organization like ours, what the risks were, and how observable and monitorable the system could be. That pilot launched in May and performed very well, involving about 1,500 business users,” Robert Purinton, chief AI officer at AdventHealth, told Becker’s.
The pilot showed value across administrative and operational teams, Mr. Purinton said, but healthcare-specific workflows required a more secure environment.
“Through our ongoing work with OpenAI, we were introduced to ChatGPT for Healthcare, which is a more secure version of the enterprise platform,” he said. “It restricts certain applications and connectors and does not connect to the live internet, ensuring healthcare data cannot be exposed externally.”
AdventHealth has since launched a pilot of ChatGPT for Healthcare focused on revenue cycle activities, such as post-discharge chart review, and support for graduate medical education, including preparing cases for residents, creating teaching aids and enabling translation into multiple languages.
“These are common, early clinical use cases across the industry, but the more secure environment gives us greater confidence to apply the technology to healthcare-specific processes,” Mr. Purinton said.
Stanford Medicine Children’s Health joined the initiative because it views OpenAI as a leader in the AI space.
“OpenAI is clearly a leader in this space today and likely will be well into the future, so we’ve been intentional about identifying partners we want to build these capabilities with over the next several years,” Keith Morse, MD, chief medical informatics officer at Stanford Medicine Children’s Health, told Becker’s. “OpenAI was high on that list.”
Although the organization is early in the collaboration, it is approaching use cases in two ways.
“We already have an internal AI chatbot that’s been available for the past couple of years. While it’s different from what we’re stepping into now, it gives us strong data on who across the organization is actively using AI day to day. Those high users are a natural place to start,” Dr. Morse said.
The second approach is focused on what Dr. Morse called “voiceless opportunities,” cross-team initiatives without a single owner but significant organizational value.
At Baylor Scott & White Health, interest in OpenAI stemmed from studying agentic AI as part of a broader shift toward a demand-driven, consumer-centered healthcare model. That interest led leaders to meet with OpenAI CEO Sam Altman to better understand the company’s healthcare ambitions, according to Nick Reddy, executive vice president and chief product officer at Baylor Scott & White.
“What stood out to us was the ability to deploy AI in a secure and responsible way. This platform operates within Baylor Scott & White’s cybersecurity posture, with all of our existing protocols and safeguards in place,” Mr. Reddy told Becker’s. “It also meets regulatory and HIPAA requirements, with the guardrails we expect as an organization. That was nonnegotiable and gave us the confidence to move forward with the partnership.”
The health system plans to roll out ChatGPT for Healthcare in its first quarter to about 2,000 to 5,000 employees, roughly 10% of its workforce, with plans to scale over time. Initial use cases include supporting workforce tasks and helping clinicians manage healthcare complexity, as well as enhancing patient and consumer experiences.
“Clinicians and operational teams are dealing with enormous volumes of information — for example, reviewing a patient record spanning more than 70 days of care. These tools can help summarize, synthesize and surface what matters most, creating significant value for our teams,” Mr. Reddy said. “The second contour is focused on patients. We strongly believe that secure, responsible AI — including agentic AI — can give people greater agency in managing their care. We see this as a catalyst for creating more differentiated, consumer-centered experiences in ways health systems haven’t been able to do before.”
Mr. Reddy added that Baylor Scott & White Health sees strong potential in ChatGPT for Healthcare but is focused on a responsible rollout.
“That means having clear success criteria, strong testing and making sure it’s working as intended,” he said. “There will be a lifecycle to work through as we scale, which is really the only factor slowing how fast and how broadly we move.”
For health systems, OpenAI’s entry into healthcare with ChatGPT for Healthcare feels like a natural evolution. The company’s models already power many tools used today, from ambient scribing to EHR tools and chatbots.
“OpenAI and ChatGPT have close to a billion monthly active users worldwide, meaning roughly one in eight people globally are already using the technology,” Mr. Reddy said. “What further differentiates OpenAI is its level of commitment to healthcare. This isn’t just a narrative — it’s backed by significant investment, amounting to several billion dollars. That combination of global scale, deep technical leadership and sustained investment in healthcare is what makes us excited about partnering with OpenAI.”
At the Becker's 11th Annual IT + Revenue Cycle Conference: The Future of AI & Digital Health, taking place September 14–17 in Chicago, healthcare executives and digital leaders from across the country will come together to explore how AI, interoperability, cybersecurity, and revenue cycle innovation are transforming care delivery, strengthening financial performance, and driving the next era of digital health. Apply for complimentary registration now.
