Europe intensifies data protection control over Google's AI model

Google's handling of personal data during the development of one of its artificial intelligence models is under scrutiny by the European data protection authority. This represents increased scrutiny of Big Tech's AI ambitions in Europe. The Irish Data Protection Commission (DPC), responsible for enforcing the European Union's General Data Protection Regulation (GDPR), has launched an official investigation into Google's Pathways Language Model 2 (PaLM 2). PaLM 2, introduced in May 2023, is the predecessor to Google's latest Gemini models, which, since December of the same year, serve as the foundation for the company's text and image generation systems. The investigation aims to determine whether Google has violated its legal obligations under the GDPR regarding the processing of personal data of EU citizens and residents of the European Economic Area. According to regulations, companies must conduct a data protection impact assessment before commencing the processing of such data, especially if this use poses a high risk to the rights and freedoms of the affected individuals. This process is particularly crucial for new technologies to ensure that the fundamental rights and freedoms of individuals are adequately considered and protected, emphasized the regulatory authority in a statement. Currently, this assessment is being examined as part of the investigation. Google has yet to respond to requests for comment. This is the latest action by the DPC against major technology companies developing extensive language models. In June, Meta paused its plans to train its Llama model using publicly available adult content on Facebook and Instagram in Europe after discussions with the Irish authority. Consequently, Meta limited the availability of some of its AI products for users in the region. A month later, X users discovered that their posts had been used without their knowledge to train systems of Elon Musk's xAI startup. The platform subsequently halted the processing of European user data for training its Grok AI model in August following legal actions by the DPC. This was the first time the authority exercised its powers to take such measures against a technology company.