Privacy Module

[PSchmiedmayer]

Use Case

Different digital health applications need to mask and automatically redact any personal information for processing, forwarding, or displaying sensitive information.

Problem

Current approaches to automatically de-identify or anonymize data often happen on web services and central infrastructure, leading to a potential risk of data being leaked or accidentally forwarded.

These use cases can include:

• Mobile health data retrieved from HealthKit
• Text input from users
• Health records received using FHIR APIs
• Documents or other data inputted into third-party APIs, including large language models.

Solution

The Spezi Privacy module should provide a set of tools that allow developers to easily de-identify personal health information retrieved in different settings.

The module should provide simple interfaces to automatically de-identify, mask, and re-identify data using local processes on the phone. Natural language processing techniques, including the NaturalLanguage framework by Apple can be used to identify key components of the provided input and provide a transparent mapping for external vendors and APIs that can be reversed or applied to responses if needed.

Additional context

Please use this issue as a discussion point for more concrete ideas about the structure of the Swift package, its focus and aim, and some first ideas around API design.

Code of Conduct

• I agree to follow this project's Code of Conduct and Contributing Guidelines