The watchdog is looking into the processing of personal data by X.
Ireland’s Data Protection Commission (DPC) has launched an inquiry into how social media platform X is using personal data pulled from public posts across the EU and EEA to educate its AI models, in particular the Grok large language models (LLMs).
Grok is a generative AI bot created by Elon Musk’s company xAI and was designed, in part, to advance conversations on X. However, it was raised early on that the vast amounts of data available on X and the often controversial subject matter could result in a chatbot that was exposed to harmful and incorrect information.
The DPC’s inquiry will examine whether X and xAI have complied with GDPR. Specifically, it will determine if personal data publicly posted was lawfully or unlawfully processed in order to train Grok LLMs. The investigation will be overseen by commissioners for data protection Dr Des Hogan and Dale Sunderland and representatives at X have been notified.
This isn’t the first time that issues have arisen with how X trains its LLMs. In July of 2024 Matthew Hodgson, CEO and co-founder of UK-based encrypted messaging platform Element, stated that Musk and X were overstepping the boundaries of digital ownership by making default the option to use public posts, interactions and conversations for further model education.
Rival social media platform Meta has faced similar issues with the Irish watchdog, however, earlier this year it was announced that AI chatbot Meta AI would launch in Europe, despite a paused roll-out stemming from “intensive discussions” with the DPC. Like X, Meta planned to train its LLMs using content pulled from materials shared on its own platforms, Facebook and Instagram.
Don’t miss out on the knowledge you need to succeed. Sign up for the Daily Brief, Silicon Republic’s digest of need-to-know sci-tech news.
