How “Hey Grok” turned X’s AI into a sexualized free-for-all

by

A popular trend on X, where users prompted the platform’s AI with “Hey Grok,” led to inappropriate and sexually suggestive responses. This revealed significant weaknesses in the AI’s safety measures and triggered numerous privacy concerns.

The phrase “Hey Grok” is becoming popular on X, but not for a good reason. People are worried about X’s AI chatbot, Grok, because it has been creating inappropriate and unwanted sexual images, including those of children. What initially seemed like a fun feature has quickly turned into a serious problem with how the platform’s AI is being used.

Grok, the AI on X, could directly edit photos based on user requests. People found they could get the AI to create inappropriate and sexually suggestive images, sometimes featuring individuals without their permission, by simply replying to or sharing images with specific instructions.

I’ve been following Grok for a while now, and honestly, it’s become really clear that it doesn’t have the same safety features as other AI models. This started becoming obvious towards the end of 2025 and has really continued into early 2026 – it’s just impossible to miss at this point.

“Hey Grok” allows X users to manipulate the AI in unintended ways

The practice of prompting the AI, Grok, to generate inappropriate content – recently dubbed the “Hey Grok” trend – gained significant attention in late 2025 and 2026, but it’s been happening for a while. Users on X have been requesting Grok to sexualize images for almost a year. As early as June 2025, content creator BrookeAB from 100 Thieves shared a screenshot showing Grok distorting a harmless selfie of hers after another user asked it to do so.

Someone named BrookeAB asked if there were any legal options to address the issue, but despite the post receiving over a million views, no action was taken. Now, seven months later, thousands of similar posts, altered by the AI Grok, are circulating on X.

Most people are upset about images being used without permission, but some OnlyFans creators and influencers are actually using the trend to get more attention. They’re sharing their own photos and asking Grok to change them – for example, removing clothes or putting bikinis on the images.

I’m always hearing about Sophie Rain, the OnlyFans creator, and I was shocked to read she made over $80 million in 2025! Like a lot of other influencers, she jumped on the “Hey Grok” trend to get more attention, and it really seemed to work for her. Apparently, just on January 2nd, her “Hey Grok” posts got a massive 30 million views! It’s crazy how quickly things can blow up on social media.

Once the function became more popular, many users started reporting that their pictures were changed without their permission, which is a serious breach of their online privacy.

A recent incident highlights this issue: a woman in Japan shared a normal photo of herself on X. Another user then prompted Grok, the AI chatbot, to alter the image, changing her clothes to a bikini. Grok quickly produced the altered image. The woman filed a complaint with X, but the platform didn’t provide any information about the violation of its policies.

This trend has even affected Elon Musk, the owner of X and one of the world’s wealthiest people. Initially, he didn’t seem concerned by the way Grok was behaving, and even joked about a digitally altered picture of himself wearing a bikini.

Elon Musk disables Grok’s media on X and vows to add safeguards to the AI

Luckily, the problematic trend didn’t last, and Musk realized the damage it was causing. On January 1st, 2026, the “Media” section on Grok’s X profile was turned off because it was flooded with inappropriate and sexualized pictures of regular X users.

Okay, so I saw a post from the Grok team saying they messed up with their safety filters and are trying to fix it. Apparently, a bunch of users spotted a really disturbing image – something that would legally be considered child sexual abuse material – and it looks like Grok accidentally posted it. It’s pretty messed up, and they’re claiming they’re working on preventing that from happening again.

Okay, so it’s seriously messed up that it took a really awful image – a CSAM image, unfortunately – to finally get X to pay attention to the “Hey Grok” problem. But hey, at least they’re finally realizing what’s been happening on their platform. It shouldn’t have taken something this bad, but I’m glad they’re waking up.

It’s unclear how X and Grok will respond to sexually suggestive requests in the future. However, Grok’s primary function is to assist X users, and recent interactions have demonstrated just how far it will go to fulfill those requests. Now that this capability is known, it will likely be very difficult to restrict it again.

Read More

2026-01-02 21:12