Andreea Lisievici

On 18 July 2024, CNIL published guidelines for organizations planning to deploy generative AI systems, emphasizing the importance of responsible and secure deployment to protect personal data. Generative AI, capable of creating various types of content, requires large datasets, often including personal data, for training. CNIL advises starting with specific needs, supervising usage, acknowledging system limitations, choosing secure and robust deployment methods, and implementing governance for GDPR compliance. Training and awareness for end users on risks and prohibited uses are crucial to mitigate potential harms and ensure proper handling of data.

On 18 July 2024, CNIL issued a Q&A document addressing the use of generative AI systems. The FAQ outlines the benefits, limitations, and compliance measures for deploying these systems, emphasizing GDPR adherence. Generative AI systems generate diverse content but pose risks such as inaccuracies and potential biases. CNIL highlights the need for proper system selection, deployment methods, risk analysis, and end-user training. The Q&A also provides guidance on compliance with the upcoming EU AI Act, which mandates transparency and accountability in AI system usage.