Ireland’s Data Protection Commission published guidance on the relationship between large language models and data protection obligations. The DPC said LLM developers may have a tendency to use vast amounts of personal data during training phases without the data subject’s knowledge or agreement and that inaccurate training data may create biases when training a model. Meanwhile, France’s data protection authority, the Commission nationale de l’informatique et des libertés, issued data protection guidance for deploying generative artificial intelligence technology.
Full story