Skip to content

PII Anonymization Knowledge Model

Overview

The PII Anonymization Knowledge Model (display name: PII Anonymization EN v#) aims at anonymizing Personal Identifiable Information (PII) contained in a text.

Automatic anonymization of PII is one of the ways to protect a person's identity when dealing with private or sensitive documents, another being pseudonymization, which can be attained with the PII Pseudonymization Knowledge Model.
The PII Pseudonymization Knowledge Model anonymises the same PII detected by the PII Knowledge Model.

Every PII is replaced by the string XXXXX.

Output structure

The model output has the same structure as any other model and is affected by the functional properties of the workflow block.
The peculiar part of the output if the modifiedText property of the extraData object, containing the anonymised text, so it's necessary to turn on the Output rules extra data functional option of the workflow block.

Example

Considering the following text:

CREDIT CARD REQUEST
NAME: John Smith
ADDRESS: 339 Rardin Dr., Broad Run, VA, 20137
BIRTHPLACE: 02/06/1992, Charlottetown, Canada
PHONE: 985-281-4501
EMAIL: [email protected]
Dear sir or madam,
I lost my credit card. Please find below the credit card's details:
Card Type: American Express 1234 4567 8901 1234 , Expiration date 09/2024, CVV 123.
This email is to request that you issue me a replacement card as soon as possible. The canceled card should be not authorized under any circumstances.
Thank you for your attention.
Sincerely

the anonymised text is the value of the modifiedText property of the extraData output object:

"extraData": {
    "modifiedText": "CREDIT CARD REQUEST \nNAME: XXXXX \nADDRESS: XXXXX \nBIRTHPLACE: XXXXX, XXXXX \nPHONE: XXXXX \nEMAIL: XXXXX \nDear sir or madam,\nI lost my credit card. Please find below the credit card's details:\nCard Type: American Express XXXXX, Expiration date XXXXX, CVV XXXXX.\nThis email is to request that you issue me a replacement card as soon as possible. The canceled card should be not authorized under any circumstances.\nThank you for your attention.\nSincerely \n"
}