How OpenAI Is Quietly Building a Healthcare Empire Around ChatGPT
When most people think of OpenAI, they think of chatbots, code generation, and the ongoing race to build artificial general intelligence. But inside the walls of one of Silicon Valley's most closely watched companies, a quieter and arguably more consequential ambition is taking shape — one that could redefine how hundreds of millions of people interact with the healthcare system. At the center of that ambition is a researcher named Karan Singhal, and the numbers surrounding his work are nothing short of staggering.
According to OpenAI, more than 230 million people now use ChatGPT for health and wellness advice every single week. That figure alone places ChatGPT in rare company, rivaling the scale of some of the world's largest health information platforms. And if Singhal has his way, that number is only the beginning.
Who Is Karan Singhal?
Karan Singhal is a leading health researcher at OpenAI whose work sits at the intersection of artificial intelligence and medicine. Unlike many AI researchers focused primarily on benchmarks and technical capabilities, Singhal is driven by a deeply human goal: making ChatGPT so reliably good at health-related guidance that it genuinely improves lives, prevents harm, and earns the trust of even the most skeptical users — including healthcare professionals themselves.
Singhal spoke exclusively with Business Insider about his work and OpenAI's broader healthcare strategy, offering a rare window into how the company thinks about one of the most sensitive and high-stakes domains any AI system can enter. His vision is expansive. He wants to help accelerate a cultural shift that he believes is already underway — one in which patients increasingly look to OpenAI's models not just for information, but as what he describes as a "protector in their care journey."
That phrase — protector in their care journey — is worth sitting with for a moment. It signals that OpenAI is not simply aiming to be a smarter search engine for symptoms. It is positioning ChatGPT as an active, ongoing participant in how people navigate their health, from understanding a diagnosis to preparing questions for a doctor to managing a chronic condition over time.
GPT-5: The First Model Built for Health From the Ground Up
One of the most significant technical developments Singhal revealed is that OpenAI's GPT-5 model family represents a fundamental shift in how the company approaches health AI. GPT-5 is the first OpenAI model family to be trained specifically, at every stage of its development, to perform better on health advice. This is not a matter of bolting on a healthcare module after the fact. It reflects a deliberate, foundational commitment to making health a core competency of the model rather than an incidental capability.
This distinction matters enormously. Earlier AI systems, including previous versions of ChatGPT, could answer health questions with varying degrees of accuracy, but they were not purpose-built to navigate the nuance, precision, and stakes that medical contexts demand. GPT-5, by contrast, has had health considerations baked into its training from the very start — shaping not just what it knows, but how it reasons, how it expresses uncertainty, and how it responds when the stakes are highest.
Why 230 Million Weekly Health Users Is Both Impressive and Sobering
The scale of ChatGPT's health usage is remarkable, but it also carries weight. When more than 230 million people per week are turning to an AI tool for health and wellness guidance, the quality and safety of that guidance becomes a matter of genuine public health importance. A well-calibrated response can help someone recognize a warning sign and seek timely care. A poorly calibrated one could do the opposite.
This is precisely why Singhal's work is so consequential. OpenAI is not operating in a low-stakes corner of the AI industry. It is building tools that real people use in real moments of vulnerability — when they are scared, confused, or simply trying to understand what is happening inside their own bodies. Getting this right is not optional. It is the entire point.
Singhal and his team are focused on embedding what they call "medical context" into ChatGPT — meaning the model should understand not just facts, but the clinical and human frameworks that shape how health information is interpreted and applied. This includes knowing when to encourage someone to see a doctor rather than rely solely on an AI's assessment, and how to communicate complex medical information in ways that are both accurate and genuinely understandable to a non-specialist.
The Road to a Google-Sized Healthcare Platform
The comparison to Google is not accidental. For years, Google has been the default destination for people with health questions — a role the search giant has occupied uneasily, given the well-documented tendency of health-related search queries to produce alarming and sometimes misleading results. ChatGPT is increasingly positioned as a more conversational, contextually aware alternative, one that can engage with the specifics of a person's situation rather than simply returning a list of links.
Whether OpenAI can achieve Google-scale dominance in the health information space depends on several factors, including continued improvements in model accuracy, the ability to earn trust from healthcare professionals and regulators, and the company's success in navigating the complex ethical and liability questions that come with AI-driven medical guidance.
What This Means for Patients, Providers, and the Future of Healthcare
For everyday users, the trajectory Singhal is charting could mean access to a knowledgeable, always-available health resource that helps them make more informed decisions, ask better questions of their doctors, and feel less alone when facing health challenges. For healthcare providers, it raises important questions about the role AI will play alongside — or in tension with — traditional clinical care.
- AI health tools could help reduce pressure on overburdened healthcare systems by handling routine informational queries and helping patients triage their own concerns more effectively.
- Improved health AI could be particularly valuable in underserved communities where access to medical professionals is limited.
- The integration of health AI into care journeys will require careful collaboration between AI developers, clinicians, regulators, and patients to ensure safety and equity.
- Trust will be the defining currency — and building it will require consistent, verifiable accuracy over time, not just impressive benchmarks.
Karan Singhal represents a new kind of figure in the AI landscape: a researcher whose work is measured not in model performance scores, but in whether the technology he builds makes people healthier, safer, and better informed. With GPT-5 designed from the ground up with health in mind, and 230 million weekly health users already relying on ChatGPT, OpenAI's healthcare ambitions are no longer hypothetical. They are unfolding in real time — and the stakes could hardly be higher.