Meddies PII: An Open Multilingual De-identification Model for Clinical Text
Researchers have introduced Meddies PII, an open-source model and dataset designed for de-identifying clinical text. The model aims to remove patient-specific information while preserving crucial clinical details necessary for AI reasoning. Meddies PII is built to handle multilingual data and various text formats found in healthcare settings, offering a starting point for hospitals needing to secure patient data for AI applications. AI
IMPACT Provides a foundational tool for healthcare AI, enabling safer use of clinical data while preserving its utility.