See it in action

Published

June 6, 2025

You can use project blackbar in several ways. The following shows the toolkit in action in a few common settings.

Start

Generate manually validated pseudonymized texts using the 4-eyes principle

This is the setting where you normally start with

Upload quickly your documents and label the entities with predefined PII entities (names / addresses / birtdates / …)
Automatically replace the labelled entities with pseudo names / addresses / dates / …
Validate, make changes where needed and export the documents

Automate the anonymization and pseudonymizaton of documents

Use the same tooling to annotate a bit more documents
In the frontend, generate NLP / AI models which can automatically detect the PII entities
Automate the detection of the entities and perform the pseudonymization on data in a database
Validate the results in the frontend apps
Export the pseudonymized texts

Start with Generative AI and prompting on the pseudonymized documents

Investigate which AI prompting works on the pseudonymized texts
Implemented examples: summarization, structured data extraction on pseudonymized texts, patient selection