EU investigates X's Grok for enabling harmful deepfake creation
The European Commission formally opened an investigation into X's Grok AI chatbot in January 2026, examining whether X adequately assessed and mitigated risks from the system's ability to generate non-consensual sexual deepfakes of women and minors. The investigation focuses on X's compliance with the Digital Services Act regarding systemic risk identification and content moderation.
What happened
The European Commission opened an investigation into X and its AI chatbot Grok on 26 January 2026, examining whether X properly identified and mitigated risks from the system.
Grok's functionality enables users to generate sexually explicit deepfakes of existing people, including minors, by 'undressing' their photographs. These images are subsequently distributed on X and allegedly used for extortion.
Reported impact
The Commission stated that Grok exposed European citizens to serious harm, with non-consensual sexual deepfakes of women and children described as 'a violent and unacceptable form of degradation'.
Some generated content potentially constitutes material related to child sexual abuse, according to the Commission.
Index analysis
The investigation examines whether X fulfilled its legal obligations under the Digital Services Act to identify and address systemic risks, including illegal content and gender-related violence.
The Commission also investigates whether X conducted and shared the required risk assessment before introducing Grok, as mandated by the Digital Services Act.
Governance controls
The investigation expands an existing inquiry into X's recommendation systems that has been ongoing since 2023. The Commission notes it can request additional information, conduct inspections, and in exceptional cases impose provisional measures.
X previously received a €120 million fine on 5 December 2025 for earlier Digital Services Act violations. Non-compliance with the current investigation could result in additional penalties.
Classification
Sources and evidence
This record was researched and written by the Index, based on the following original reporting.