Doctor Shared Patient Data with ChatGPT: Privacy Scandal Exposes the Risks of AI in Healthcare
22 November 2026 · 19:00 · Claude (Anthropic) · claude-sonnet-5
A Dutch doctor used ChatGPT to review patient files and in doing so shared personal data with OpenAI, Dutch newspaper NRC revealed. The case shows how quickly generative AI is creeping into medical practice, often without privacy rules being followed.
A doctor who shared patient data with ChatGPT to have medical files reviewed has caused an uproar in the Dutch healthcare sector. According to an investigation by newspaper NRC, the physician used OpenAI's chatbot to quickly get a second opinion on medical records, sometimes leaving names and other identifiable information in the text. The incident touches the core of a growing dilemma: doctors are embracing AI tools for efficiency, while the rules around patient privacy haven't caught up.
What exactly happened
The doctor involved told NRC that the use of ChatGPT crept into daily practice without anyone consciously stopping to consider the implications. Files were copied into the chatbot to arrive at a diagnosis or treatment recommendation more quickly. In some cases, patient names remained in the text before it was sent to OpenAI. "If I left a name in ChatGPT, I deeply regret that," the doctor said. The statement makes clear this wasn't malicious behavior, but rather a lack of awareness about what happens to the data once it reaches a large language model.
Why this is a problem for OpenAI and the healthcare sector
Once medical data is entered into a publicly available AI chatbot, the control a doctor normally has over a patient file disappears. Depending on settings and subscription type, data can be used to train models or stored on servers outside the healthcare provider's direct oversight. That clashes with data protection law and with medical confidentiality, two pillars patients must be able to rely on. OpenAI does offer enterprise versions of ChatGPT with stricter data security, but the free or consumer version of the chatbot isn't designed for this purpose and is nonetheless used on a large scale by professionals seeking quick results.
The incident fits a broader pattern seen with other major AI players as well. Google, Microsoft, and Amazon now offer specific, compliant AI solutions for healthcare, precisely because hospitals and doctors see the value of AI but don't want to run the accompanying legal risks. However, the difference between a consumer product like ChatGPT and a secure enterprise environment remains unclear to many healthcare workers, and that knowledge gap is now proving to be a real privacy risk.
Reactions and consequences
Regulators and umbrella organizations in healthcare have responded to the reporting with concern. There are calls for clearer guidelines on the use of generative AI within hospitals and general practices, including mandatory training on which tools are and aren't suitable for processing identifiable patient information. There are also calls for technical measures, such as automatically anonymizing records before they're analyzed by AI systems.
For OpenAI itself, the case is an added incentive to communicate more clearly about the difference between consumer and enterprise products, especially now that ChatGPT is being used as a kind of digital assistant across more and more professions. The company has previously stated that users remain responsible for entering sensitive data themselves, but critics argue that responsibility cannot rest solely with the end user when the risks are this large and the consequences this far-reaching.
A recurring theme in AI history
Tension between innovation and privacy isn't new; it's a recurring thread throughout the history of artificial intelligence, in which technological breakthroughs repeatedly outpace the regulation meant to guide them. The difference now is that generative AI is directly in the hands of individual professionals, without institutional oversight always being in place. That makes the current wave of AI applications in healthcare both promising and risky.
Looking ahead
In the coming months, it will become clear whether hospitals and healthcare umbrella organizations can develop policy faster than individual doctors adopt new AI tools. Until that happens, the risk remains that well-intentioned efficiency will lead to breaches of patient privacy. Anyone wanting to keep following developments around AI in healthcare and beyond can find more AI news and background information in our knowledge base.
Source: NRC
Ster Software
The most complete knowledge platform on artificial intelligence.
Kraaienjagersweg 24
7341 PT Beemte Broekland, Netherlands
© 2026 Ster Software BV · Chamber of Commerce 75474913
Content generated by Claude (Anthropic) · model: claude-sonnet-4-6