The Irish Information Safety Fee (DPC) has introduced that it has commenced a “Cross-Border statutory inquiry” into Google’s foundational synthetic intelligence (AI) mannequin to find out whether or not the tech big has adhered to information safety rules within the area when processing the non-public information of European customers.
“The statutory inquiry issues the query of whether or not Google has complied with any obligations that it could have needed to undertake an evaluation, pursuant to Article 35[2] of the Common Information Safety Regulation (Information Safety Affect Evaluation), previous to partaking within the processing of the non-public information of E.U./E.E.A. information topics related to the event of its foundational AI mannequin, Pathways Language Mannequin 2 (PaLM 2),” the DPC stated.
PaLM 2 is Google’s state-of-the-art language mannequin with improved multilingual, reasoning, and coding capabilities. It was unveiled by the corporate in Could 2023.
With Google’s European headquarters primarily based in Dublin, the DPC acts as the first regulator accountable for ensuring the corporate abides by the bloc’s stringent information privateness rulebook.
The DPC stated an inquiry is essential to make sure that people’ basic rights and freedoms are safeguarded, particularly when processing of such information when growing AI techniques can result in a “excessive threat.”
The event comes weeks after social media platform X completely agreed to not prepare its AI chatbot, Grok, utilizing the non-public information it collected from European customers with out acquiring prior consent. Again in August, the DPC stated X consented to droop its “processing of the non-public information contained within the public posts of X’s E.U./E.E.A. customers which it processed between 7 Could 2024 and 1 August 2024.”
Meta, which not too long ago admitted to scraping each Australian grownup Fb person’s public information to coach its Llama AI fashions with out giving them an opt-out, has paused its plans to make use of content material posted by European customers following a request from the DPC over privateness issues. It has additionally suspended the usage of generative AI (GenAI) in Brazil after the nation’s information safety authority issued a preliminary ban objecting to its new privateness coverage.
Final yr, Italy’s information privateness regulator additionally quickly banned OpenAI’s ChatGPT due to issues that its practices are in violation of knowledge safety legal guidelines within the area.
