An AI-specific data protection impact assessment is a structured evaluation of how an AI system may affect personal privacy and which safeguards could reduce those effects. It belongs within governance across data collection, curation, training, and model outputs, where risks can accumulate without clear oversight. section 3.18
A useful assessment considers:
- Whether data was collected with consent and clear provenance. section 3.18
- Whether curation introduces or amplifies bias through inadequate documentation, balance, or review. section 3.18
- Whether training transformations hinder tracing or deleting specific personal information. section 3.18
- Whether outputs reveal private details or generate biased content, and how technical, legal, policy, ethics, and affected-community perspectives should respond. section 3.18
