This paper describes a framework for privacy-preserving data sharing, addressing technical challenges as well as some data sharing issues more broadly.
The paper builds on the 2017 ACS paper, Data Sharing Frameworks, expanding the concept of a Personal Information Factor and introducing a Data Safety Factor with recommendations for threshold settings.
The paper speaks to some of the challenges of trusted data sharing. These include concerns with the implications of data quality, use of outputs, the changing risk inherent in the release of results over time, and the need to develop a ‘social licence to operate’.
This paper further develops the concept of a quantified ‘Five Safes’ data analytics framework and briefly examines the implications of such frameworks when artificially intelligent algorithms are used to analyse data. The paper provides a set of recommendations to trial the data sharing framework within the context of developing a national information governance framework.