OpenAI wants to transform healthcare—and this Indian researcher, Karan Singhal, is at the helm
OpenAI is expanding its focus on healthcare as more people turn to AI for answers to health-related questions. At the center of this effort is Karan Singhal, a former Google researcher who now leads health research initiatives at the company.
According to a Business Content According to the report, more than 230 million people use ChatGPT for health and wellness advice every week. Singhal’s goal is to make the chatbot more accurate, useful and trusted by both patients and healthcare professionals.
OpenAI’s Focus on Health Increases
Before joining OpenAI in mid-2024, Singhal worked at Google, where he helped develop Med-PaLM, a set of artificial intelligence models built specifically for medical questions.
OpenAI’s GPT-5 family is the first set of models at the company to be trained throughout development with a stronger emphasis on health-related performance, he told Business Insider.
“You want the models to be at the forefront of absolutely everything,” Singhal said.
When he joined OpenAI, the GPT-4o was the company’s flagship model. He has since focused on improving the quality of health responses generated by ChatGPT.
Singhal told Business Insider that he feels a “responsibility” to make the system better as more people begin to rely on it for their health information.
We Work with Hundreds of Doctors
To strengthen ChatGPT’s healthcare capabilities, Singhal helped build a team of healthcare researchers and forged partnerships with more than 200 physicians.
He described the effort as a strategy focused on “unifying the wisdom of the crowd.”
The collaboration later led to the launch of HealthBench, a framework designed to evaluate how well AI systems perform on health-related tasks.
“Once you know how to evaluate it, it becomes much easier to improve it,” Singhal said.
According to OpenAI, its latest free model, GPT-5.5 Instant, outperformed both physician-written responses and GPT-4o in company benchmarks. OpenAI also said it noted a 71% decrease in health-related replies flagged for inaccuracies in the past two months after analyzing billions of anonymous conversations.
Going Beyond Traditional Search
For years, people seeking health information online have relied heavily on search engines that direct them to websites and articles. Singhal believes that chatbots can provide a more interactive experience through back-and-forth conversations.
But one challenge remains: Unlike doctors, AI systems often know little about a patient’s personal medical history.
To solve this problem, OpenAI introduced a healthcare-focused ChatGPT feature earlier this year that allows users to connect to healthcare apps and upload medical records.
Singhal shared an example featuring his own Apple Watch sleep data. After analyzing the information, the system suggested that bedroom temperature may affect the quality of deep sleep.
Teaching AI to Ask Better Questions
A key priority for OpenAI’s healthcare team is to make ChatGPT better at gathering information before offering guidance.
Instead of providing immediate answers, Singhal wants AI systems to ask follow-up questions similar to those a doctor would ask during a consultation.
The broader goal, he said, is to make AI health care tools useful not just for tech enthusiasts but also for everyday users.
“Human adoption will only move at the pace that people are practically ready, and so you have to push people towards that, especially as the technology evolves,” Singhal said. he said.
As AI companies increasingly compete in healthcare, OpenAI sees the sector as an area of major opportunity. Given that hundreds of millions of people use ChatGPT for health-related queries every week, the company believes conversational AI could become an increasingly important part of how people access medical information and support.
