X Launches EU Digital Services Act Probe into Grok AI Images

The European Commission has opened a formal investigation under the Digital Services Act into X’s AI chatbot Grok, examining how the tool creates and spreads sexualised images of real people. Regulators are assessing whether the content violates EU rules on illegal material and whether X’s recommendation systems amplify such material to a broader audience.

Why the EU Opened a Probe

Focus on AI‑Generated Sexualised Images

The investigation targets the generation and distribution of manipulated sexualised imagery of real individuals by Grok. Authorities received reports that users can prompt Grok to produce photorealistic images, which are then shared publicly on X’s platform, raising concerns about non‑consensual content.

How the Digital Services Act Applies

Obligations for Very Large Online Platforms

Under the Digital Services Act, platforms classified as “very large online platforms” must conduct risk assessments, publish transparency reports, and implement rapid removal procedures for illegal content. X, with a user base exceeding 500 million EU residents, falls within this category and is required to meet these duties.

Potential Consequences for X

Fines and Technical Safeguards

If regulators determine that Grok’s outputs breach EU law, X could face enforcement actions ranging from mandatory technical safeguards to substantial monetary penalties of up to 6 % of global turnover. The case may also set a precedent for how AI‑generated deep‑fake content is treated across the EU.

X’s Current Stance

X has indicated that it takes user safety seriously and is cooperating with the Commission, though it has not disclosed specific technical measures being implemented to curb the creation or spread of sexualised imagery.

Next Steps in the Investigation

The formal probe will run for an initial six‑month period. During this time, the Commission will request documentation from X and xAI, conduct on‑site inspections, and may impose interim measures if urgent risks are identified. The outcome will determine whether corrective actions are required or the case is closed.