AI Transparency

This page summarises a document that is executed in Polish. The Polish version is the only binding one; this summary has no legal effect and is provided for orientation. Questions in English are welcome — write to us and we will answer them.

What the document covers

• Where AI is used — transcribing the visit, laying out the note, matching dictionary codes to what was said, filling official forms from material you supply, and answering questions grounded in cited sources.
• What AI does not do — it does not diagnose, does not recommend treatment and does not decide anything. Nothing reaches the medical record without the physician's approval.
• Verifiability — answers point to their source: the transcript passage, the page of the paper, the section of the Summary of Product Characteristics.
• Your data and models — data entrusted to us is not used to train models, neither by us nor by our providers.
• Limits — language models make mistakes. The interface is built so that every generated item is reviewed before use, and so that the review is easy rather than nominal.