The Irish Data Protection Commission has opened an inquiry into whether Alphabet Inc.'s Google complied with European Union privacy laws in the development of its artificial intelligence model.
The DPC said that the probe will look at whether Google carried out a data protection impact assessment, as required by the bloc's sweeping General Data Protection Regulation, before it processed EU residents' personal data used in its Pathways Language Model 2 (PaLM 2) foundational model.
"A data protection impact assessment, where required, is of crucial importance in ensuring that the fundamental rights and freedoms of individuals are adequately considered and protected when processing of personal data is likely to result in a high risk," the DPC said in a statement Thursday.
Big Tech's influence over AI has been drawing increasing scrutiny from regulators around the globe, amid concerns they are using access to personal data to dominate the emerging sector. The EU has passed a number of laws that place guardrails to limit it's activities in the bloc, including the world's most extensive AI rules.
Ireland's DPC is responsible for overseeing the data protection practices of companies like Google that have their EU headquarters in the country.
A spokesperson for Google in Ireland didn't immediately respond to a request for comment.
Earlier this month, Elon Musk's X Corp. agreed to stop processing the personal information of European users to train its AI chatbot Grok. It followed the DPC's filing of a request with Ireland's High Court to stop X from processing the data for use in the company's AI model, citing EU data protection restrictions and alleging that X's action risked jeopardizing users' rights.
The DPC also asked the EU-level European Data Protection Board to start a discussion on the interplay between data protection and AI model training – an issue that has in recent months given rise to disputes between data watchdogs and tech companies, including X and Meta Platforms, Inc.
In June, Meta said it would delay the launch of the Meta AI chatbot in Europe after the DPC requested the company pause its plan to train its large language models with posts from users there.
