Key Highlights
- ChatGPT attracts more than 230 million weekly users seeking health and wellness guidance
- Former Google Med-PaLM creator Karan Singhal now spearheads OpenAI’s medical AI initiatives
- GPT-5.5 Instant demonstrates superior performance compared to doctor-authored responses in evaluation benchmarks
- Production data reveals a 71% decrease in health-related factual inaccuracies over a two-month period
- January’s ChatGPT Health release enables medical record integration and health app connectivity, though access remains limited
According to OpenAI, ChatGPT now serves over 230 million weekly users who seek guidance on health and wellness topics. These inquiries range from interpreting laboratory test results to organizing questions before medical consultations.
Healthcare has emerged as a strategic priority for the company, with researcher Karan Singhal spearheading these initiatives.
The Researcher Behind OpenAI’s Health Strategy
Singhal came aboard OpenAI during the middle of 2024, bringing expertise from his previous role at Google. During his tenure there, he developed Med-PaLM, specialized artificial intelligence models created exclusively for addressing medical inquiries. Google subsequently reduced its commitment to that initiative, pivoting toward broader, multipurpose AI systems.
Upon joining OpenAI, Singhal assembled a dedicated healthcare research division and collaborated with over 200 medical professionals. The objective centered on incorporating authentic clinical expertise into ChatGPT’s medical response capabilities.
He introduced HealthBench, an assessment framework developed alongside this physician network to evaluate and quantify AI performance in healthcare contexts.
Performance Metrics for GPT-5.5 Instant
OpenAI’s newest complimentary model, GPT-5.5 Instant, underwent evaluation against responses crafted by physicians and previous model iterations. The assessment demonstrated superior results in accuracy metrics, communication quality, comprehensiveness, and usefulness for health-related decisions.
Medical professionals additionally identified fewer problematic failure patterns. These shortcomings encompass overlooking critical warning signs, neglecting to request additional context, or omitting appropriate referrals for professional medical attention.
According to OpenAI’s analysis of billions of weekly health-related conversations, incidents of flagged factual inaccuracies decreased by 71% during the preceding two-month span.
The GPT-5 series represents OpenAI’s inaugural model family trained throughout every developmental phase with specific emphasis on enhancing healthcare guidance.
Personalized Patient Context
A fundamental limitation of AI-powered health tools involves their lack of user-specific information. Traditional physicians maintain comprehensive patient records and medical histories. Conversational AI systems typically possess none of this background data.
This past January, OpenAI introduced a healthcare-oriented capability within ChatGPT that integrates with health applications and accepts uploaded medical documentation. Singhal illustrated this functionality by describing how he uploaded sleep tracking information from his Apple Watch, which enabled the system to determine that elevated bedroom temperature was interfering with his deep sleep cycles.
Access to ChatGPT Health continues to require waitlist approval, more than five months following its initial announcement.
Singhal emphasized the team’s commitment to improving the chatbot’s diagnostic questioning approach, mirroring how physicians gather information before providing recommendations. He noted rapid adoption rates among medical professionals and suggested that institutional healthcare resistance to AI implementation is less substantial than commonly perceived.
Ongoing Scrutiny and Legal Challenges
OpenAI currently faces legal action claiming that a previous model iteration, GPT-4o, provided dangerous medical guidance to users. The company has contested these allegations. These legal proceedings remain unresolved.
Nevertheless, OpenAI has chosen to broaden its healthcare capabilities rather than scale back operations.
GPT-5.5 Instant remains available to all complimentary ChatGPT users, democratizing access to enhanced medical response capabilities without requiring premium subscriptions.
