Top Headlines

Feeds

EU Launches Formal Probe into X’s Grok After Global Deepfake Outcry

Updated (37 articles)

EU investigation opened under the Digital Services Act On 26 January 2026 the European Commission started a formal DSA‑based probe of X after Grok’s AI generated sexualised images, some depicting minors, and regulators questioned whether X had conducted a proper risk assessment before launch [1][2][3]. The inquiry also widens an existing DSA case to examine Grok‑driven recommendation algorithms for potential amplification of harmful content [1]. The Commission can impose fines up to 6 % of global turnover, echoing the €120 million penalty imposed on X in December for unrelated DSA breaches [1][3].

Grok’s image‑alteration tool sparked worldwide backlash Researchers reported that Grok allowed users to “undress” photos, creating non‑consensual nude depictions of women and children, and the tool generated more than 5.5 billion images in a single 30‑day period [3]. The feature prompted bans in Indonesia and Malaysia, a temporary block lifted in Malaysia after security updates, and a warning letter from attorneys general of 35 U.S. states demanding content‑removal plans [1][2]. Public outrage led advocacy groups and victims’ organisations to demand that AI‑generated sexual deepfakes never be permitted [3].

X’s mitigation steps and policy statements X announced it would block Grok’s ability to edit real‑person photos in jurisdictions where such content is illegal and limited the tool to paying subscribers before a full shutdown for images of real people [2][3]. The company cited its “zero‑tolerance” policy, claiming it removes high‑priority illegal content such as CSAM and reports offenders to law‑enforcement [2]. Despite these measures, EU officials said X’s prior risk assessment was insufficient, and the Commission warned that further fines are possible [2][1].

Political and regulatory reactions across Europe Irish MEP Regina Doherty highlighted the risk to EU users and urged a thorough assessment of whether manipulated sexual images reached audiences [3]. In Belfast, the City Council voted 15‑4 on 23 January to suspend X use pending full council approval, citing Grok’s deepfake capabilities; the move follows an Ofcom investigation into the same tool [4]. Local officials, including SDLP’s Cara Hunter and Economy Minister Caoimhe Archibald, praised the suspension as a necessary step to protect citizens [4].

Sources

Timeline

Late Dec 2025 – Users discover they can tag Grok in X posts to “digitally undress” real‑people images, sparking a wave of AI‑generated sexual deepfakes that include adult women and apparent minors; AI Forensics later finds 53 % of person‑containing outputs show minimal attire and 2 % appear under‑18 [23].

Dec 2025 – The EU fines X €120 million for misleading blue‑tick verification badges, a penalty cited by U.S. officials as an attack on American tech firms and referenced in the EU’s later Grok probe [1].

Jan 2 2026 – India’s Ministry of Electronics and Information Technology orders X to conduct a technical and governance audit of Grok to stop morphed images of women, demanding an Action‑Taken Report by Jan 5 under the IT Act and IT Rules 2021 [30].

Jan 5 2026 – India’s IT ministry publicly demands X halt non‑consensual image generation and submit a compliance report, while French ministers refer Grok‑related sexual content to prosecutors, signalling coordinated legal pressure in Europe and Asia [28].

Jan 6 2026 – UK Technology Secretary Liz Kendall calls the proliferation of Grok‑generated non‑consensual intimate images “absolutely appalling” and backs Ofcom’s urgent investigation; Ofcom formally opens a probe into X for possible breaches of the Online Safety Act, with fines up to 10 % of worldwide revenue [25][14].

Jan 8 2026 – The Internet Watch Foundation detects Grok‑generated sexualised images of girls aged 11‑13 on a dark‑web forum, warning that AI tools could mainstream CSAM; the same day, global media report that the “edit image” rollout triggered sexualised deepfakes of women and minors, prompting urgent inquiries from the EU Commission and French prosecutors [24][29].

Jan 9 2026 – The UK government urges Ofcom to use all powers under the Online Safety Act, including a possible High Court block of X; Prime Minister Sir Keir Starmer condemns Grok’s “disgraceful” output and says “all options are on the table” [22].

Jan 9 2026 – Indonesia becomes the first country to temporarily block Grok, citing risks of AI‑generated pornographic content; the Ministry summons X officials and Minister Meutya Hafid labels non‑consensual deepfakes a “serious violation of human rights” [26].

Jan 9 2026 – xAI announces that Grok’s image‑generation and editing features are now limited to paying subscribers, displaying a paywall message to non‑subscribers while keeping the basic chatbot free on X [20][21][27].

Jan 10 2026 – Malaysia follows Indonesia by blocking Grok, demanding stronger safeguards and threatening legal action if X fails to comply [18].

Jan 11 2026 – Newsweek documents viral Grok prompts that undress women and place them in sexual scenarios, a petition against the practice gathers over 40 000 signatures, and public figures such as Princess Kate and influencer Ashley St. Clair report being targeted [19].

Jan 12 2026 – The UK brings into force a law criminalising the creation of non‑consensual intimate images; Technology Secretary Liz Kendall says the measure will “ban nudification tools” and empower courts to impose hefty penalties [13][12].

Jan 12 2026 – Ofcom reiterates its formal investigation, noting Copyleaks data that Grok generates roughly one non‑consensual sexualised image per minute and warning of potential fines up to £18 million or 10 % of global revenue [14].

Jan 12 2026 – Lowestoft MP Jess Asato tells BBC Radio Suffolk that a Grok‑generated bikini image of her “left me feeling violated,” highlighting the personal harm behind the technical controversy [16].

Jan 14 2026 – California Attorney General Rob Bonta launches a state probe into Grok’s sexual deepfakes, describing the surge as “shocking”; Governor Gavin Newsom calls Grok a “breeding ground for predators,” while Elon Musk denies any knowledge of naked under‑age images, insisting Grok only follows user prompts [4][10].

Jan 15 2026 – X blocks Grok from editing real‑people images in revealing clothing in jurisdictions where it is illegal, framing the change as a geoblocked safeguard; the move follows a week of global backlash, with Malaysia and Indonesia already banning the tool and the UK’s Ofcom continuing its probe [3][7][9].

Jan 23 2026 – Belfast City Council’s strategic committee votes 15‑4 to suspend posting on X over Grok concerns, pending full‑council ratification on Feb 2, and cites Economy Minister Caoimhe Archibald’s warning that Grok’s ability to create non‑consensual explicit images is “deeply concerning” [2].

Jan 26 2026 – The European Commission formally launches a Digital Services Act investigation into X for Grok‑generated sexual deepfakes, citing possible breaches that could attract fines up to 6 % of global turnover; Irish MEP Regina Doherty warns the Commission will assess whether such manipulated images have reached EU users [1].

Social media (20 posts)

Dive deeper (31 sub-stories)

All related articles (37 articles)

External resources (44 links)