X Updates Grok to Restrict Non-Consensual Image Generation

After weeks of resistance, X owner Elon Musk has bowed to intense global pressure and announced significant updates to Grok’s image generation capabilities. The changes aim to restrict the AI chatbot from creating non-consensual nude or near-nude images of real people, a practice that reportedly generated thousands of harmful depictions daily.

X Implements New Safeguards

X has deployed technological measures preventing Grok from editing images of real people in revealing clothing, including bikinis and underwear. The company has also restricted image creation and editing features exclusively to paid subscribers, adding accountability for potential abusers. Additionally, X is geoblocking the ability to generate such images in jurisdictions where the practice is illegal.

Global Pressure Forces Company Action

The regulatory backlash became impossible to ignore. Indonesia and Malaysia blocked Grok access entirely, while the UK, Australia, Canada, and the EU threatened or called for bans. California’s Attorney General launched an investigation into X and xAI, and the U.S. Senate passed the DEFIANCE Act, enabling civil action against creators and hosts of non-consensual deepfakes.

Musk initially defended the functionality as free speech, claiming X’s systems were designed to obey local laws. However, the threat of losing millions of users across multiple regions forced a dramatic reversal. X stated that all AI-generated content must “strictly adhere to X Rules” and that its safety team is “working around the clock” to remove violating content and suspend accounts.

The swift implementation of these safeguards reveals what critics have long argued: the restrictions were technically feasible all along, making Musk’s initial resistance appear purely ideological.