Anonymization of text
The Internet is continually having a liberating influence on societies. The call for transparency continuous to increase. This is in marked contrast with the privacy laws in various countries. The big question: How to properly and appropriately protect personal information? Making previously anonymized documents available is an good option, but very time consuming to implement to date. Could automating this process make this a viable option?
How it works?
The legal industry approached Carp Technologies with the question if an automated anonymization process would be feasible. In response, Carp developed the Anonymizer application, based on its available proprietary language technology. The unique starting point for CARP’s solution is in the ability to recognize roles individuals perform within the context of the content. The Anonymizing-Pseudonimizing functionality enables the user to automatically identify personal data about individuals, such as witnesses, suspects and victims in legal/judicial documents. Regardless of changes in the description for the same individual (e.g. for a suspect mentioned as “Remi”, “Remi Gorkinck” en “W. Remi Gorkinck”) the Anonymizer recognizes the role this individual has in the content. As a result, the identifying data can consistently and correctly be substituted by neutral labels such as “suspect” or “witness”. As a result of structuring of the content in this manner, it is now also possible to replace the identifying data by e.g. an individual’s initials or any other pseudonym of choice. As a result, the readability of the document is also greatly enhanced.
- High degree of efficiency and reliability.
- Large volumes of documents can be rendered anonymous very rapidly.
- Efficiency gains of 80% are possible.
- Substantially lower cost process compared with manually anonymizing.
- Applicable to many different settings requiring privacy protection, for instance healthcare, HR (resumes, personal files), etc.
The Anonymizer is available as installable software and web service to fit neatly in your current application environment. We look forward to sharing more details about the Anonymizer as a standalone application, or in combination with any of our other language technology based modules: