Privacy-Preserving AI

Privacy-preserving AI refers to techniques that let organizations train, run, or analyze AI systems while limiting how much personal or confidential data is exposed along the way. Common approaches include differential privacy, which adds controlled statistical noise so that no single individual's record can be identified from a model's outputs, and federated learning, which trains models across distributed data sources without centralizing the raw data. For compliance and risk teams, these methods matter because they can help satisfy data minimization and purpose limitation duties under laws such as the EU General Data Protection Regulation (GDPR), though they do not automatically make a system compliant on their own.

1 item