Fig. 1: Threats to data privacy.
From: Addressing contemporary threats in anonymised healthcare data using privacy engineering
a Who is involved? Contemporary privacy threats come from intended and unintended users. b What are the Risks? To learn an individual’s characteristics or identity. This can occur by Inferring characteristics from anonymised data including membership of a specific group (membership inference attacks, MIA), inferring a sensitive attribute or inferring identity. These extend risks of direct data leakage of PII. c What are the solutions? Such threats could be addressed by Health Data Privacy Frameworks, that guide the selection of privacy engineering and include components such as the precise task, the type of user, the type of data required, the duration of time for which data are needed, and ethical considerations.
