Irish Data Protection Regulator to Investigate Google AI

Eire’s knowledge safety authorities have launched a probe into Google’s AI mannequin, and whether or not it complies with GDPR.

The Irish Information Safety Fee (DPC), An Coimisiún um Chosaint Sonraí, is the EU’s lead privateness regulator for Google. The DPC has opened a cross-border statutory inquiry into Google Eire, below Part 110 of the Information Safety Act 2018.

The regulator is investigating whether or not Google ought to have carried out a Information Safety Affect Evaluation (DPIA) earlier than processing knowledge belonging to EU and EEA residents in its foundational AI mannequin, often called Pathways Language Mannequin 2 (PaLM 2).

Cross-border processing is the place private knowledge is processed in multiple member state or is more likely to have an effect on residents in multiple member state.

Eire’s DPC identified that, below Article 35 of the GDPR, knowledge processors ought to perform an influence evaluation when knowledge processing “is more likely to end in a excessive danger to the rights and freedoms of people.”

The regulator added that that is particularly vital the place new applied sciences are concerned. The method wants to make sure knowledge processing “is critical and proportionate,” and that organizations have put safeguards in place.

Earlier this month, the DPC ended a courtroom case towards X (previously Twitter) after the social media agency mentioned it might restrict the way it makes use of private knowledge from EU residents to coach its AI fashions.

In June, Meta – which owns Fb and Instagram – delayed giant language mannequin (LLM) coaching primarily based on public content material from the 2 social media channels, once more following a request by Eire’s DPC. This has delayed the European launch of Meta’s AI assistant, Meta AI.

The social media agency dedicated to working with the DPC, in addition to the UK’s Info Commissioner’s Workplace (ICO), to handle privateness considerations forward of future AI mannequin coaching.

Learn extra about safety for LLMs: OWASP Releases Safety Guidelines for Generative AI Deployment

Picture credit score: Shaheerrr / Shutterstock.com