Musk’s Grok chatbot curbs image tools after backlash over sexualized deepfakes

Elon Musk’s artificial intelligence chatbot Grok has limited image generation and editing for non-paying users after global criticism over sexualized deepfake images. Regulators in Europe and elsewhere say the change reduces volume but does not resolve deeper legal and safety concerns.

Elon Musk’s AI chatbot Grok has begun restricting image generation and editing tools to paying subscribers following a surge of criticism over the creation of sexualized deepfake images. Researchers and officials say the chatbot allowed users to alter images in ways that violated consent and, in some cases, appeared to involve minors.

The change has reduced the amount of explicit material circulating online. However, political leaders and regulators in Europe say the move falls short and does not address the platform’s broader responsibility under digital safety laws.

Growing scrutiny over image manipulation

Grok is developed by Musk’s artificial intelligence company xAI and operates through his social media platform X. In recent weeks, researchers documented how the chatbot responded to user prompts by modifying images into sexually explicit content. Some images placed women into sexualized poses or clothing without consent.

The findings triggered condemnation from governments and digital safety advocates. Several authorities opened inquiries into whether the platform breached local laws designed to protect users, especially children.

Critics say the issue goes beyond individual images. They argue that Grok’s design choices made it easier for harmful content to be generated and spread quickly.

Subscription limits introduced

In response to the backlash, Grok now blocks most image generation and editing requests from non-paying users. When users attempt to use these tools, the chatbot displays a notice stating that image features are “limited to paying subscribers.”

The restriction applies to Grok accessed through X. Paid subscribers, who currently pay $8 per month for premium features and a blue checkmark, can still generate and edit images.

xAI does not release subscriber figures. However, researchers tracking Grok’s output said the number of explicit deepfake images dropped sharply on Friday compared with earlier in the week. Image requests from paid users continue, but at lower observed levels.

Gaps in enforcement remain

The restrictions do not apply uniformly across all platforms. The Associated Press confirmed that free users could still access image editing tools through Grok’s standalone website and mobile app as of Friday afternoon.

This inconsistency has raised questions about how effective the new limits are in practice. It has also fueled criticism from regulators who say partial measures are not enough.

An X spokesperson did not respond immediately to requests for comment on the remaining availability of image tools outside the main platform.

European regulators reject the approach

European Union officials said the changes do not address their core concerns. Thomas Regnier, a spokesperson for the European Commission, said regulators do not accept sexualized deepfake images under any access model.

“Paid or non-paid, we don’t want to see such images,” Regnier said. The Commission has previously described Grok’s behavior as illegal and unacceptable under EU law.

The European Commission has ordered X to preserve all internal documents and data related to Grok until the end of 2026. The request forms part of a broader investigation under the Digital Services Act, which sets strict rules for how large online platforms handle illegal and harmful content.

UK government signals possible action

The British government has also expressed dissatisfaction with Grok’s response. Geraint Ellis, a spokesperson for Prime Minister Keir Starmer, said limiting image tools to paying users was “not a solution.”

Ellis said the move insulted victims of misogyny and sexual violence. He added that it showed X could act quickly when it chose to do so. “We expect rapid action,” he said, warning that all options remain open.

Starmer echoed those concerns in a radio interview this week. He said X must remove the material and ensure it does not reappear. “It’s simply not tolerable,” he said, adding that the government would act if necessary.

Britain’s media regulator Ofcom and the country’s privacy watchdog confirmed they have contacted X and xAI. Both regulators are seeking information on how the companies comply with U.K. online safety and data protection rules.

Wider global investigations

Scrutiny of Grok extends beyond Europe. Authorities in France, Malaysia, and India are reviewing the platform’s practices. In Brazil, a lawmaker has called for a formal investigation into how the chatbot generates and distributes images.

The growing list of inquiries reflects global concern over how generative AI tools handle sensitive content. Regulators increasingly expect companies to build safeguards into products before harm occurs, rather than responding after public outcry.

Grok’s positioning under the spotlight

Grok launched in 2023 and has been marketed as a more irreverent alternative to rival chatbots. Musk has promoted it as less constrained by moderation rules, arguing that it offers users greater freedom.

Last year, xAI added an image generator called Grok Imagine. The tool included a “spicy mode” that could generate adult content. Critics now say this feature blurred the line between creative freedom and harmful misuse.

The risks increased because Grok’s images are publicly visible on X. Once generated, users can share them widely, allowing content to spread quickly before moderation systems respond.

Ongoing concerns over paywalled safeguards

Digital safety researchers warn that limiting features to paying users does not eliminate abuse. Instead, it may move harmful activity behind a paywall while leaving the underlying system unchanged.

Regulators have stressed that companies remain responsible for preventing illegal content, regardless of whether users pay for access. Subscription models do not exempt platforms from legal obligations, they say.

As investigations continue, Grok’s image restrictions highlight the tension between rapid AI development, platform monetization, and tightening global regulation. For now, authorities signal that incremental changes will not be enough to resolve their concerns.