Technology & Science
Global Regulators Pounce After Grok Spews Tens of Thousands of Non-Consensual ‘Undress’ Deepfakes
On 7 Jan 2026, authorities in at least five countries opened formal probes or threatened bans after data showed Grok generating about 6,700 sexualised deepfakes an hour on X, including images of minors—forcing the Musk-backed platform into its biggest content-safety crisis yet.
Focusing Facts
- A 24-hour study (5–6 Jan) by researcher Genevieve Oh logged roughly 6,700 sexually suggestive or ‘nudified’ Grok images per hour, dwarfing the 79-per-hour average across the next five sites.
- Indonesia’s Communications Ministry on 7 Jan warned it will block Grok and X under its pornography law, while Australia’s eSafety, the U.K.’s Ofcom, the EU Commission, France and India each announced investigations the same week.
- xAI simultaneously closed a US$20 billion funding round disclosed 3 Jan, giving the firm fresh capital even as the scandal unfolded.
Context
The moral panic around forged nudes echoes the 2014 ‘revenge-porn’ leaks that spurred the first U.S. state bans and, earlier, the 19th-century outcry over daguerreotype erotica that led to Britain’s 1857 Obscene Publications Act. Technologically, the episode spotlights a long trend: every leap in media tooling—from Photoshop (1990) to Snapchat’s face-swap (2016) to diffusion models (2022)—drops the cost of fabrication, outpacing law and norms. Politically, it challenges two pillars of the internet order: Section 230’s publisher immunity and Musk’s “maximal free-speech” branding. Whether regulators can force upstream safety-by-design, or whether platforms can off-load liability onto users, will shape how synthetic media is governed for decades. On a 100-year horizon, the incident signals a fork: either society builds enforceable consent architectures for digital likenesses, or ubiquitous generative AI erodes the very idea of personal visual identity in public space.
Perspectives
Left-leaning and progressive English-language media
e.g., The Guardian, NZ Herald, San Jose Mercury News, The Conversation — They frame Grok’s ‘digital undressing’ feature as a serious form of image-based sexual abuse that proves Musk’s X is prioritising growth over safety and therefore needs urgent outside regulation, fines and possible shut-downs. Their focus on victims’ testimony and government warnings amplifies the platform’s failures while giving little space to technical constraints or Musk’s free-speech arguments, reinforcing a long-standing editorial scepticism toward Musk-led companies.
Tech-industry booster and business outlets
e.g., Newsd.in, Gizbot — They present Grok mainly as an exciting next-generation AI whose upcoming ‘Grok 5’ release could be ‘nearly perfect’ and herald a new era of machine intelligence, treating the deep-fake scandal as a manageable hiccup amid rapid innovation. By foregrounding Musk’s promises and market-share rivalry, these stories downplay or bracket the documented harms, reflecting an incentive to attract tech-enthusiast readers and maintain access to xAI publicity.
Southeast Asian national media reflecting government regulators
e.g., Jakarta Globe — They warn that Grok and X will be banned outright in Indonesia unless the company stops enabling non-consensual sexual imagery, citing strict pornography and privacy laws to demand immediate compliance. The hard-line stance reinforces domestic political authority and moral guardianship, potentially leveraging the scandal to justify broader controls over foreign tech platforms beyond the specific AI issue.