OpenAI Dives into Healthcare, Hires Top Execs to Build AI Medical Tools

The expansion aims to develop tools that can improve clinical decision-making, support patient engagement, and strengthen healthcare delivery.

OpenAI, the company behind ChatGPT, is stepping into the healthcare space with fresh leadership appointments and a shift in strategy.

The company, which had so far relied on external collaborations, is now moving to build its own AI-driven health applications.

The expansion aims to develop tools that can improve clinical decision-making, support patient engagement, and strengthen healthcare delivery.

To drive this expansion, the company has made two key leadership appointments. Daniel Etra, formerly with Instagram, has been named Head of Healthcare Product Strategy.

He will lead efforts to design and scale AI tools that directly support clinicians and patients.

Ashley Alexander, previously with Doximity, has been appointed Vice President of Health Products. She will oversee the development and implementation of OpenAI’s health-focused solutions.

These appointments come after a series of healthcare-focused initiatives this year. Earlier this year, in May 2025, OpenAI launched HealthBench, an open-source benchmark developed with input from 262 physicians across 60 countries.

The dataset includes 5,000 simulated medical interactions and 48,000 evaluation criteria covering areas such as emergencies, global health, and diagnostic interpretation. It is being used to assess whether AI models provide accurate and safe responses in clinical settings.

Following the release of HealthBench, OpenAI rolled out GPT-5, its most advanced model. GPT-5 demonstrated measurable improvements in factual accuracy and adherence to clinical instructions compared with earlier versions.

OpenAI has also tested clinical applications. In partnership with Penda Health, a primary care network in Nairobi, Kenya, the company deployed AI Consult, a GPT-4o-powered assistant. The tool was evaluated across nearly 40,000 patient visits involving 20,000 clinicians.

Results showed reductions in diagnostic errors by 16 percent, treatment errors by 13 percent, and history-taking errors by 32 percent. The system included a feedback mechanism with green, yellow, and red alerts, and no cases of harm were reported.

Moreover, in the U.S., OpenAI has engaged with healthcare organizations for operational applications. Oscar Health is using OpenAI models to streamline administrative processes such as claims documentation, aiming to reduce manual workloads.

Furthermore, Color Health has also integrated OpenAI’s technology into oncology workflows, cutting cancer treatment planning times from weeks to minutes while keeping physicians in control of decisions.

Stay tuned for more such updates on Digital Health News