Irish privacy regulator Data Protection Commission (DPC) is investigating whether Google properly protected the personal data of EU citizens before using this data to train its foundation model, PaLM2.
The DPC typically acts as an EU entity in privacy cases involving mostly the big U.S. tech companies because they often have European headquarters in Ireland. This investigation focuses mainly on the personal data of EU citizens. Reuters writes that Google’s parent Alphabet allegedly used this data to train the foundation AI model PaLM2. In the investigation, the privacy watchdog wants to determine to what extent the tech giant properly protected this data before it was used in the training process.
Earlier process against Grok
The DPC is investigating the data training processes of AI providers more common. Just last week, X stated to the privacy regulator that it will not train its Grok AI models using data from EU citizens until they have granted permission via an opt-in option. This follows an earlier action by the privacy watchdog.
Last summer, the privacy regulator released a set of guidelines specifically for training AI/LLM models that developers must adhere to in order to be compliant with various European laws and regulations.
Also read: Complaints filed in European countries against AI training Grok on X-posts