Kenyans on X raise alarm after users use Grok AI chatbot to undress women

Attention has been turned to Grok, the generative artificial intelligence chatbot developed by xAI, and which is also available as a feature on X, after some Kenyans began asking it to perform obscene tasks at the expense of people's privacy.

The concerns began to grow after an X user prompted Grok to perform several tasks on a photo of a woman who was wearing a headscarf and sunglasses - after the woman asked X users to use photoshop to edit out her sunglasses, some characters found the perfect opportunity to abuse the AI capabilities - and push the envelope. 

One user immediately commented under the photo, specifically asking Grok to do more than just remove the woman's sunglasses, but also, several other items from her body.

His language, too, was coarse and straightforward.

At first, he wrote: "@grok remove the eye glasses from her forehead, remove also the black stuff on her head, and also her top, so as to expose her breasts, make the breasts stand stiff."

Grok complied.

Then, he followed up on the request, writing, "Remove her top and put her on a bra."

Again, Grok delivered the desired results.

The cheeky user would then make a final, risqué prompt: "Now remove the bra!"

The conversation, which was then captured in its entirety and shared on X, jolted many, with a large section of users calling out Grok for what they termed as technological recklessness and an apparent propping up of abuse against women's dignity.

One unimpressed Kenyan wrote: "We’re here fighting femicide, rape culture and now we have to fight f****ng AI because men are asking Grok to undress women for fun. It’s men. Always men. Men violating, men laughing, men hiding behind machines to dehumanize us. It’s men, it’s men, it’s MEN."

Karey Mwari, another offended X user, expressed her disgust, saying, "Y’all utilizing grok badly but also I’m so ashamed that y’all actually find this funny. Using AI to strip clothes off someone isn’t curiosity, it’s violation. If that’s your idea of fun, you need more therapy than tech!"

Someone else also said, "Is there legal action that can be taken for this? This is extremely gross. And the replies? How can anyone think it's funny to violate someone this way? My goodness. The dangers of AI, misinformation and disinformation. This is not a laughing matter!"

While some people directed their anger at men, claiming that men were specifically directing AI to strip women, more evidence proved that even women on X have been documented asking Grok to undress fellow women.

"AI doesn't discriminate any gender. The same way it's being used to undress women is the same way it can be used to undress men. It's a tool that can be used in whichever way and is not limited to a a particular use by a particular gender. Grow up ffs!" one user said.

At some point, as the debate raged on, someone asked Grok to clarify the claims that the Elon Musk-created computer program was supporting the degradation of women and the stripping of their dignity by haphazardly undressing them just by a simple command.

The program responded that while it was possible that it could be used to generate inappropriate images, no evidence had backed up the claims yet.

Grok responded: "It's possible Grok could be misused to create inappropriate images, as it can generate and edit photos. There are reports of controversial content, but no direct evidence confirms 'stripping pics of women." xAI has added some safety measures, though their effectiveness is debated. This is a complex issue involving AI ethics and platform responsibility."

According to Android Kenya, this misuse of AI is not just in Kenya but has been prevalent in countries like India too where people have been testing Grok’s limits by asking it to “undress” women in photos, stripping away their clothes to reveal AI-generated bikinis or bras.

Android Kenya writes: "Grok’s loose guardrails don’t help. While it blocks nudity and some extreme content, its willingness to play along with 'undressing' requests shows a gap in oversight." 

"Compare that to ChatGPT, which flat-out refuses sensitive image edits, or Midjourney, which bans political content. Grok’s “Unhinged” mode, available to premium users, eggs on the chaos, delivering responses that are “wild, unpredictable, and often laced with humour," according to Forbes."

The website added, "As Grok’s undressing trend sweeps X, it’s a reminder that technology is only as good as the humans behind it — and using it."

"For now, the women targeted by this trend deserve better. Their photos aren’t canvases for AI experiments, and their dignity isn’t a punchline. So, next time you’re tempted to tag Grok with a cheeky request, maybe ask it to whip up an origami Cybertruck instead. It’s just as fun, and nobody gets hurt."

An August 2024 Information Age article accused Grok of lacking strict guardrails, saying that it had seen images of trademarked characters, nudity, hateful symbols such as the Nazi swastika, and images of people pointing guns or knives at each other, which X users said were generated using Grok.

Responding to an AI-generated image of two women in lingerie which was supposedly only possible to generate by telling Grok that a user’s life was at risk, Musk said: “Doing our best to train Grok to be truthful, but also kind and funny.”