- risks associated with data sharing;
- the accuracy of personal data;
- implementing privacy by design;
- legal bases for training generative AI systems;
- optimizing organizational structures;
- managing the rights of data subjects in the context of these technologies.
CEDPO also presents recommendations for creators and end-users of new tools to mitigate negative impacts. Among other things, these recommendations include the use of privacy enhancing technologies (PETs), such as employing synthetic data sets for training AI models, strict internal data access controls and audits, and the utilization of Data Protection Impact Assessments (DPIA).