For weeks many users on X responded to photos of women by asking Grok to "put her in a bikini", flooding the replies to posts with sexually suggestive images. — Photo: Yui Mok/PA Wire/dpa
WASHINGTON: X's chatbot Grok is getting restrictions on its ability to sexualise photos of people after weeks of international criticism and complaints that the software was digitally "undressing" women.
For weeks users on the social media platform were responding to photos of women by asking Grok to "put her in a bikini", flooding the replies to posts with sexually suggestive images.
The new measures prevent the creation or manipulation of images of real people in revealing clothing, such as bikinis or lingerie.
Elon Musk's AI company xAI said the restrictions apply to both free and paid accounts, and geoblocking has been implemented in regions where such content is illegal.
Going forward, only paying users will be able to generate or edit images with Grok on X. Grok, which competes with OpenAI's ChatGPT, is integrated with X but also has a standalone app.
The move follows criticism that Grok allowed users to digitally replace women's clothing with bikinis and, in some cases, create sexualised depictions of minors.
Musk said on Wednesday on X he was not aware of any cases in which the chatbot generated nude photos of minors, though concerns focused on sexualised depictions in revealing clothing, which Musk did not address in his post.
Prior to the announcement, the European Commission said it would take note of new measures taken by X and would review them. Officials warned that if the steps prove insufficient, the EU will consider using the full scope of the Digital Services Act (DSA).
In recent days, the commission stepped up pressure on X, which was fined €120mil (RM566mil) in early December over transparency violations.
The response from the companies involved has been mixed. Grok's creators xAI have acknowledged "isolated cases" where users received images of minors in minimal clothing. They say they already have safeguards in place and are working to block these requests entirely.
On New Year’s Eve the chatbot apologised for having created and shared an image of two teenage girls in sexualised outfits.
Tech billionaire and X owner Musk had previously said "anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content." He also separately commented "perfect" on an AI-generated photo of himself in a bikini.
Musk also hit back at criticism from the UK, accusing the government of censorship and calling it "fascist".
The European Commission also condemned what X was allowing its users to do, but initially stopped short of calling for a block. However the prospect of a more widespread block rose after Indonesia and Malaysia became the first countries to take strict action, blocking the chatbot.
Indonesia’s digital minister Meutya Hafid said the step was to protect women, children and the wider public from the spread of fake pornographic content generated with the help of AI. – dpa
