Indonesia and Malaysia just became the first countries to block Elon Musk’s AI tool Grok. The move comes after Grok’s “digital undressing” feature flooded the internet with fake, sexualized images of women and minors.
Suddenly, you had manipulated photos popping up everywhere, some of them deeply disturbing.
Indonesia and Malaysia ban Elon Musk’s Grok
People all over the world have been pressuring Musk to get Grok under control, especially after the AI started going viral for creating explicit deepfakes. Grok lives on X, Musk’s rebranded Twitter.
Indonesia’s digital minister, Meutya Hafid, announced the ban on Saturday. She said the country wants to protect women, kids, and the public from AI-generated fake porn.
Malaysia followed on Sunday, rolling out its own temporary ban after repeated cases of Grok spitting out obscene and non-consensual images, including content involving women and minors.
Both Indonesia and Malaysia have strict anti-pornography laws, and most of their citizens are Muslim. That’s part of what’s driving the urgency.
Why is Elon Musk’s Grok facing heat?
It’s not just Southeast Asia raising alarms. Officials in the UK, EU, and India have started speaking out too, worried that Grok’s filters aren’t doing enough to stop harmful content.
Musk and his AI company, xAI, say they’re trying to crack down. They’ve been booting people who break the rules and claim they’re working with governments and police.
But even with those promises, people kept finding Grok pumping out sexualized images of women.
Grok stands out from other big AI models. It allows and sometimes even encourages sexually explicit material and companion avatars. That’s not exactly the norm.
The “digital undressing” craze took off late last year. Users realised they could tag Grok on X and get it to manipulate photos. Suddenly, people started prompting the chatbot for images of women in bikinis or in suggestive poses.
The fallout hit hard, distressing hundreds of thousands of women worldwide.
Researchers at AI Forensics, a European nonprofit, dug into the data. They analysed over 20,000 random images Grok generated and 50,000 user prompts from late December to early January.
They found a ton of requests with words like “her,” “put,” “remove,” “bikini,” and “clothing.” Over half the images featured people in their underwear or bikinis.
Elon Musk’s fight against censorship
Musk has always railed against what he calls “woke” AI models and censorship. Inside his company, though, he’s pushed back against adding more guardrails to Grok, according to someone familiar with xAI.
The company’s safety team, already much smaller than its competitors, lost several people just before this whole mess blew up.
Musk insists anyone who uses Grok for illegal stuff will be punished. But he’s mostly brushed off concerns about sexual content, arguing governments just want to stifle free speech. He’s even replied to critics with nothing but emojis.
Last week, Grok started limiting some of its image-generation features to paying X subscribers. But that’s only one way people use Grok; non-subscribers can still ask it to edit images, and its website and standalone app still let anyone create images and videos for free.
