Elon Musk’s artificial intelligence tool Grok is facing increased scrutiny after the European Union launched a formal investigation into its image-generation features.
The probe follows widespread backlash in several countries, including Australia, over Grok allowing the creation of sexualised images of women and children without their consent. The European Commission said it is examining whether X, formerly Twitter, adequately assessed and mitigated the risks linked to the rollout of Grok.
The investigation will focus on potential violations of the EU’s Digital Services Act (DSA), particularly the risk of disseminating illegal content such as manipulated sexually explicit images, including material that may amount to child sexual abuse content. The Commission said these risks “appear to have materialised,” exposing EU citizens to serious harm.
EU Executive Vice-President for Tech Sovereignty, Security and Democracy Henna Virkkunen described sexual deepfakes as an unacceptable and harmful practice, stressing that the investigation would determine whether X had met its legal obligations or failed to protect the rights of women and children.
Concerns about Grok emerged late last year, prompting regulators and governments in Australia, the UK, France, Poland, India, Malaysia and Brazil to contact X. Australia’s eSafety Commissioner said it had seen a recent increase in reports related to Grok generating sexualised or exploitative imagery, particularly involving children, and warned it would use its enforcement powers where necessary.
The European Commission noted that it had already taken action in 2025 against so-called “nudify” services used to create AI-generated child exploitation material.
X owner Elon Musk has denied that Grok generates illegal images, saying the system only responds to user prompts and is designed to comply with the laws of each country. He acknowledged that attempts to exploit the system through adversarial prompts may occur but said any such issues are fixed immediately.
The eSafety Commission has not confirmed whether X has formally responded to its concerns.
