European regulator opens GDPR probe into Google’s AI model
European regulators have launched an inquiry into Google’s AI large language model and whether processing personal data of European citizens to train the tool has breached the bloc’s impact assessment rules under the GDPR.
As Google’s lead regulator in Europe, Ireland’s Data Protection Commission (DPC) will oversee the statutory inquiry into the company’s Pathways Language Model 2 (PaLM 2). The DPC said this is part of its wider work alongside European peers to regulate the processing of personal data in the development of AI models and systems.
The DPC said the GDPR requires regulated firms to undertake data protection impact assessments before they begin to process personal data in the EU that is likely to result in high risk.
The probe follows growing regulatory pressure on social media platforms’ use of personal data that they hold on Europeans to train new AI models.
Last month, X suspended training for its AI-powered chatbot Grok to sidestep court action initiated by the DPC. In July, Meta pulled its AI large language model Llama from Europe, citing the “unpredictable nature of the European regulatory environment”.
Meta previously delayed the launch of AI chatbot Meta AI after the DPC intervened and asked the company to stop training its models using posts from users of its Facebook and Instagram platforms in potential breach of the GDPR.
