Google hit with European privacy probe over its AI system

The Irish privacy watchdog opened an investigation Thursday into one of Google’s artificial intelligence systems, marking a new step in European regulators’ escalating fight with tech firms over the development of new technology.

The Irish Data Protection Commission said Google may not have evaluated how and if its gathering of Europeans’ personal data to develop one of its AI models could threaten fundamental rights and freedoms.

The potential violation of the European Union’s landmark privacy rulebook, the General Data Protection Regulation (GDPR), may result in a multimillion-euro fine of up to 4 percent of Google’s global annual revenues.

The probe comes as Europe’s privacy regulators are cracking down on tech firms’ development of their AI systems over data protection concerns.

The Irish authority, which oversees many Big Tech firms that have their EU headquarters in the country, in the past year has ordered X, Meta and Google to postpone their AI training and stop using some Europeans’ public data, like posts.

Earlier this month, X agreed to stop using European users’ public posts collected without their consent after a court battle. Meta announced in June that it had suspended its plan to gather Europeans’ public posts and photos over concerns from regulators. The firm later said it would not release one of its new AI product citing the bloc’s “unpredictable” regulations.

Google last year also had to delay the launch of its generative AI Gemini, formerly known as Bard. The Irish authority is now scrutinizing Pathways Language Model 2 (PaLM 2), a model used to generate text in different languages and code.

“We take seriously our obligations under the GDPR and will work constructively with the DPC to answer their questions,” said Alex McPhillips, a Google spokesperson.