Developing and validating an anonymization copilot

Training artificial intelligence and machine learning (AIML) models requires a large amount of data. However, access to data for AIML projects has been problematic in practice, largely due to privacy concerns. One approach to enable data sharing and data access is to anonymize the data. However, in the Canadian context, a national standard and a unified definition of concepts such as anonymization, de-identification, and non-personal information do not currently exist. This makes it difficult for privacy professionals to determine what is the appropriate anonymization process to follow and has resulted in a shortage of professionals across the country with the expertise to perform effective anonymization.
The purpose of this project is to use the GPT-4 large language model (LLM) to develop and validate a customized Canadian anonymization copilot to guide professionals through the anonymization process. The use of LLMs to support professionals is a research area that is showing a lot of promise. This copilot would enable privacy professionals to understand how to anonymize data and would generate ready-to-use anonymization plans for them.

Faculty Supervisor:

Khaled El El Emam

Student:

Partner:

TELUS (Toronto, ON)

Discipline:

Computer science

Sector:

Information and cultural industries

University:

University of Ottawa

Program: