This investigation adds to the numerous investigations already underway in several countries on the subject.
The California Attorney General announced the opening of an investigation into xAI due to the online proliferation of images of naked women and children generated with Grok, an artificial intelligence (AI) assistant created by Elon Musk’s company.
“xAI appears to facilitate the large-scale production of non-consensual deepfakes used to harass women and girls online, including through the social network X”, highlighted, in a statement, Democratic Attorney General Rob Bonta.
This investigation adds to the numerous investigations already underway in several countries on the subject.
In the same note, the prosecutor urged xAI “to take immediate measures” to ensure that the production of these images “never happens again”.
“We have zero tolerance for the creation and distribution, through artificial intelligence, of non-consensual intimate images or child pornography,” he added.
In recent weeks, international outrage has grown around Grok and its ability to modify images, especially those published on the social network X, also owned by Elon Musk.
Some users took advantage of this tool to publish or respond to messages containing photos of women, accompanied by instructions such as “dress her in a bikini” to receive a hyper-realistic edition (“deepfake”, in English) in return.
Since January 9th, the image creation functionality has been disabled for free users of X, which claims to be taking steps to remove this illegal content and suspend the accounts involved.
The California prosecutor also accused xAI of advertising its ability to undress people “as a marketing tool, which predictably led to the proliferation of sexualized content without the consent of the individuals involved.”
Last week, an analysis conducted by non-governmental organization (NGO) AI Forensics of more than 20,000 images generated by Grok revealed that more than half depicted scantily clad individuals, 81% of which were women and 2% of which appeared to be minors.
Also today, Malaysia joined several other countries that are reviewing the functioning of the AI tool, including France, the United Kingdom and Indonesia.
The British communications regulator, Ofcom, opened a formal investigation against
Indonesia announced on Saturday a restriction on access to the ‘chatbot’ due to the creation of “indecent, extremely offensive and non-consensual images, including content involving women and minors”.
