Recent Regulatory Developments in Training Artificial Intelligence Models Under the GDPR

By: Marcus Evans & Rosie Nance (Norton Rose Fulbright)
In 2024, many organizations are increasingly exploring how to leverage their data to launch or enhance AI programs. A common approach is to use this data to train AI models or refine third-party models. Often, this involves processing personal data, requiring compliance with the General Data Protection Regulation (GDPR) in the EU or UK.
X, formerly known as Twitter, recently attempted to train its AI chatbot, Grok, using data from user posts. However, following discussions with Ireland’s Data Protection Commission (DPC), the company agreed to suspend this data processing. This article examines the key events and offers considerations for data controllers when using personal data to train AI models.
On May 7, 2024, X introduced a new setting under ‘Settings and Privacy’ regarding Grok’s training. This setting, enabled by default, allowed users to opt out. Users were asked if they would “Allow your posts as well as your interactions, inputs, and results with Grok to be used for training and fine-tuning.”
Details on how to opt out are available on X’s website, which also notes that setting an account to private will prevent the use of posts for training Grok (only public posts are used).
In its privacy notice, X states: “We may use the information we collect and publicly available information to help train our machine learning or artificial intelligence models for the purposes outlined in this policy.” The company appears to rely on legitimate interests (Article 6(1)(f) of the GDPR) to process data for model training and provides further details on its legitimate interests assessment on its website.
The Irish DPC applied to the Irish High Court to compel X to suspend data processing for this purpose. The case was heard on August 8, 2024. X’s Global Government Affairs department has noted that the DPC’s order would affect all AI models at X, not just Grok…
Featured News
FTC v. Meta Trial Turns to Market Definition
Apr 28, 2025 by
CPI
Marriott to Acquire CitizenM for $355 Million, Expanding Urban Lifestyle Offerings
Apr 28, 2025 by
CPI
Thomson Reuters Urges Third Circuit to Block Ross Intelligence’s Copyright Appeal
Apr 28, 2025 by
CPI
Merck KGaA to Acquire SpringWorks for $3.9 Billion
Apr 28, 2025 by
CPI
Federal Judge Dismisses Mario Chalmers’ Antitrust Lawsuit Against NCAA Over NIL Rights
Apr 28, 2025 by
CPI
Antitrust Mix by CPI
Antitrust Chronicle® – Mergers in Digital Markets
Apr 21, 2025 by
CPI
Catching a Killer? Six “Genetic Markers” to Assess Nascent Competitor Acquisitions
Apr 21, 2025 by
John Taladay & Christine Ryu-Naya
Digital Decoded: Is There More Scope for Digital Mergers In 2025?
Apr 21, 2025 by
Colin Raftery, Michele Davis, Sarah Jensen & Martin Dickson
AI In the Mix – An Ever-Evolving Approach to Jurisdiction Over Digital Mergers in Europe
Apr 21, 2025 by
Ingrid Vandenborre & Ketevan Zukakishvili
Antitrust Enforcement Errors Due to a Failure to Understand Organizational Capabilities and Dynamic Competition
Apr 21, 2025 by
Magdalena Kuyterink & David J. Teece