Grok faces global scrutiny for sexualised AI deepfakes

Governments and regulators around the world are cracking down on sexually explicit content generated by Elon Musk’s xAI chatbot Grok on X, launching probes, imposing bans and demanding safeguards in a growing global push to curb illegal material.

Here are some reactions from governments and regulators since the start of January:

EUROPE

The European Commission on January 26 opened an investigation into whether Grok disseminates illegal content such as manipulated sexualised images in the EU. The probe will examine whether X properly assessed and mitigated risks as required under the bloc’s digital rules.

The Commission had on January 8 extended an order sent to X last year to retain and preserve all internal documents and data related to Grok until the end of 2026.

Ireland’s Data Protection Commission opened an investigation into Grok on February 17, examining its handling of personal data and potential to generate harmful sexualized images and videos, including of minors. The DPC oversees X in the EU as the company’s European headquarters are in Ireland.

Spain’s government meanwhile ordered prosecutors to probe X, Meta META.O and TikTok over alleged distribution of AI-generated child sexual abuse material, Prime Minister Pedro Sanchez said.

Britain’s media regulator Ofcom has launched an investigation into X to determine whether sexually intimate deepfakes produced by Grok violated its duty to protect people in the UK from content that could be illegal, under the country’s Online Safety Act framework.

Paris prosecutor’s cybercrime unit raided X’s office in Paris on February 3 and ordered Musk to face questions in April regarding a widening investigation over alleged algorithmic bias, complicity in the detention and diffusion of images of child‑pornographic nature and the violation of a person’s image rights with sexually explicit deepfakes.

Italy’s data protection authority warned that using AI tools to create “undressed” deepfake imagery of real people without consent could amount to serious privacy violations and, in some cases, criminal offences.

ASIA

India’s IT ministry sent X a formal notice on January 2 over alleged Grok-enabled creation or sharing of obscene sexualised images, directing the content to be taken down and requiring a report on the actions being taken within 72 hours.

Japan also probed X over Grok, stating that the government would consider every possible option to prevent the generation of inappropriate images.

Indonesia’s communications and digital ministry said it had blocked access to Grok, a move digital minister Meutya Hafid said was meant to protect women and children from AI-generated fake pornographic content, citing Indonesia’s strict anti‑pornography laws.

Malaysia restored access to Grok for its users after X implemented additional safety measures, its communications regulator said on January 23.

The Philippines will reinstate access to Grok after its developer pledged to remove image-manipulation tools that had sparked child-safety concerns, the country’s cybercrime investigation unit said on January 21.

AMERICAS

California’s governor and attorney general said on January 14 they were demanding answers from xAI amid the spread of non-consensual sexual images on the platform.

Canada’s privacy watchdog said it was widening an existing investigation into X after reports that Grok was generating non-consensual, sexually explicit deepfakes.

Brazil’s government and federal prosecutors gave xAI 30 days to prevent the chatbot from spreading fake sexualised content, according to a joint statement on January 20.

OCEANIA

Australia’s online-safety regulator eSafety said on January 7 it was investigating Grok-generated sexualised deepfake images, assessing adult material under its image‑based abuse scheme and noting current child-related examples it had reviewed did not meet the legal threshold for child sexual abuse material under Australian law.

HOW HAS xAI RESPONDED?

xAI said on January 14 it had restricted image editing for Grok AI users and blocked users, based on their location, from generating images of people in revealing clothing in “jurisdictions where it’s illegal”. It did not identify the countries.

It had earlier limited the use of Grok’s image generation and editing features only to paying subscribers.