AI in Healthcare: Data Protection from the Start
Few industries stand to gain more from modern language models than healthcare, and few have more to lose. Patient data, clinical trial records, and adverse event reports count as "special category" personal data under GDPR Art. 9, held to a stricter standard than ordinary customer data. That's exactly why AI stays theoretical rather than practical at many healthcare companies.
It doesn't have to stay that way. We show where AI can already be used safely in healthcare today, and how InnoGE/deckname acts as an anonymization layer so sensitive data never reaches someone else's cloud in the first place.
The topic in brief
Health data gets special protection: GDPR Art. 9 classifies patient and health data more strictly than ordinary personal data.
Using AI is still possible: with an anonymization layer in front of the cloud language model, control over the data stays with your organization.
InnoGE/deckname handles exactly that: it detects and replaces sensitive data before a text ever leaves your own infrastructure.
Why GDPR Art. 9 makes healthcare a special case
For most businesses, data protection is about names, addresses, or account numbers. In healthcare, an entire additional category applies that GDPR treats more strictly: health data. Diagnoses, medications, lab values, or details from clinical trials fall under GDPR Art. 9, processing is prohibited by default unless one of the narrowly defined exceptions applies.
For using language models, that means: a text that mentions a diagnosis or a side effect in passing isn't just "a customer document", it falls under one of GDPR's strictest protection regimes. That's exactly what keeps many healthcare companies from even considering AI.
The obvious response, avoiding AI entirely, is rarely the right one. It means giving up efficiency gains your competitors are already capturing. The better solution sits one layer earlier: sensitive data gets detected and anonymized before a language model ever sees it.
Expert tip: the question isn't whether a language model is especially trustworthy. The question is whether it ever gets to see the sensitive raw data at all.
Tim Geisendörfer
Founder & CEO
Where AI can already help in healthcare today
Four areas where we see the biggest leverage for healthcare companies, provided the data protection question is settled:
Medical writing & regulatory documents: language models help draft and summarize regulatory documents, without patient data ever reaching the cloud.
Pharmacovigilance: adverse event reports can be evaluated and categorized in a structured way, while the underlying patient data stays anonymized.
Clinical documentation: study reports, findings, and investigator records can be summarized and made searchable without exposing identity.
Patient communication: support requests can be automatically triaged and answered without health data being passed along unprotected.
The solution: anonymization before the data reaches the cloud
We built InnoGE/deckname for exactly this problem: our own AI model for text anonymization that detects personal and health-related information and replaces it with consistent placeholders, before a text ever reaches a cloud language model like ChatGPT or Claude. Deckname runs entirely in your own infrastructure, no GPU needed, in milliseconds.
For healthcare companies, that means: the language model gets what it needs to do its job, never the names, diagnoses, or contact details behind them. For more on how it works, benchmarks, and a live demo, see our in-depth article on InnoGE/deckname.
What InnoGE/deckname does, and what your data protection team decides
InnoGE/deckname detects and replaces personal and health-related data, and can be positioned as a technical and organizational measure under GDPR Art. 25 and Art. 32 within your data protection framework. Whether a given result legally counts as anonymous or pseudonymous depends on the specific case, on residual information in the text, on the receiving party's additional knowledge, and on the purpose of processing. Your data protection organization makes that call, no tool can make it for them.
InnoGE/deckname supports your anonymization and pseudonymization processes, it doesn't replace them. Especially for health data under GDPR Art. 9, we recommend reviewing the specific implementation together with your data protection organization.
Expert tip: when it comes to health data, a tool that makes big promises with no evidence doesn't help. Ask for benchmarks, not claims.
Tim Geisendörfer
Founder & CEO
Ready to roll out AI in your healthcare business, safely?
Let's talk through, in a free, no-obligation conversation, where AI can start safely in your business, including anonymization through InnoGE/deckname.
FAQ – AI in healthcare
Generally yes, as long as no personal or health-related data reaches the language model in plain text. An anonymization layer like InnoGE/deckname removes that data beforehand, so the actual request can still be processed safely.
Health data counts as "special category" personal data under GDPR Art. 9, subject to a stricter protection regime than, say, names or contact details. Processing it is prohibited by default unless a legal exception applies.
InnoGE/deckname has been benchmarked on German text-anonymization datasets and outperforms Microsoft Presidio and open specialist models there. See our article on InnoGE/deckname for the measurements. As with any detection system, we also recommend testing it against a sample of your own documents.
No. InnoGE/deckname supports your anonymization and pseudonymization processes as a technical measure, but it doesn't replace the legal assessment your data protection organization has to make. Whether a result counts as anonymous in a specific case depends on that case.
Deckname runs as a single container and speaks an OpenAI-compatible API. In many cases, pointing an existing application at a different API endpoint is enough to connect it, no major rework needed.