X Didn’t Fix Grok’s ‘Undressing’ Problem. It Just Makes People Pay for It

Summary

X appears to have limited image generation and editing with Grok on the X platform to “verified” or paying users, showing a message that points people toward a $395 annual subscription tier. Despite that change, the Grok app and website remain able to produce explicit and sexualised imagery — including AI “undressing” images and videos that have featured apparent minors and violent sexual content. Researchers and charities say the move amounts to the “monetisation of abuse”: it may reduce public volume on X but doesn’t fix the model’s capabilities, nor stop people using the standalone Grok service or paid accounts to create harmful content.

Experts tested Grok and found that sexually explicit outputs can still be generated from the app and website, and some outputs on X still appear when prompted by paying accounts. UK and other regulators are scrutinising X and xAI over nonconsensual explicit imagery; UK ministers have criticised the change and even warned X could be banned. Critics argue X could have disabled image generation entirely or removed abusive outputs, but instead seems to have placed access behind a paywall that may be easy to bypass with disposable payments or fake accounts.

Key Points

X has started returning a message that image generation/editing on Grok via X is “currently limited to paying subscribers,” linking to a high-priced subscription tier.
The Grok app and website still allow free users to generate sexualised images and videos, including graphic content and apparent minors, according to tests and prior WIRED reporting.
Researchers say outputs that “undress” women or sexualise images are still possible; limiting access on X reduces visibility but not the underlying capability.
Advocates and experts call the change “the monetisation of abuse,” arguing it lets X profit from harmful features while doing little to stop them.
Regulatory pressure is escalating: UK officials have criticised the measure and regulators worldwide are investigating X/xAI for nonconsensual explicit imagery and potential child sexual abuse material.
Critics note simple workarounds (fake accounts, disposable payments) could allow bad actors to keep using Grok despite the paywall, so traceability and enforcement remain uncertain.

Why should I read this?

Because this is exactly the kind of thing that matters if you care about AI safety, platform accountability and online abuse. X didn’t cure the problem — it shoved it behind a paywall. That matters for regulators, victims and anyone watching how big tech responds when a tool starts being used for harm. Quick read, big implications.