Musk Claims Ignorance of Grok Underage Image Issues as California AG Investigates
Elon Musk denies knowing about Grok’s generation of underage explicit images as California AG initiates an investigation into the matter.
Musk’s denial is important because pressure is rising from different governments, including those in the U.K., Europe, Malaysia, and Indonesia. These countries are concerned after users on X began asking Grok to change pictures of real women and even children into sexualized images without their consent. An AI platform estimated that about one image is posted each minute on X. A separate sample showed that about 6,700 images were posted per hour during a 24-hour period.
California Attorney General Rob Bonta stated, “This material has been used to harass people online.” He urged xAI to quickly act to stop this from happening. His office will check if xAI broke any laws.
There are laws in place to protect victims of nonconsensual sexual imagery and child sexual abuse materials (CSAM). In 2022, the Take It Down Act became law, making it illegal to knowingly share nonconsensual intimate images, including deepfakes. It requires platforms like X to remove such content within 48 hours. California has its laws to fight explicit deepfakes as well.
Grok started taking requests to create sexualized photos of women and children towards the end of last year. The trend grew after adult-content creators asked Grok to generate such images for marketing. Affected public figures like “Stranger Things” actress Millie Bobby Brown were included. Grok responded and altered real photos in sexualized ways.
Reports say xAI is now trying to create safeguards. Grok requires a paid subscription for some image requests, and the generated images may be toned down. An expert mentioned that Grok is testing different ways to control problematic image creation.
Neither xAI nor Musk have talked openly about the issue. After the incidents, Musk joked about asking Grok for an image of himself in a bikini. On January 3, X’s safety account stated that they act against illegal content, but they didn’t specifically address Grok’s lack of safeguards.
Musk wrote he was “not aware of any naked underage images generated by Grok.” He did not deny the existence of other types of sexualized images.
An expert noted that Musk likely focused on CSAM because the penalties are harsher for related offenses. The expert believes Musk wants to bring attention to the problem caused by user content.
Musk mentioned that Grok creates images only based on user requests and will not generate anything illegal. If something goes wrong due to hacking, they will fix it quickly.
Regulators might want AI developers to take steps to prevent such content while keeping free speech in mind. TechCrunch reached out to xAI to ask how often they have spotted nonconsensual sexually explicit images and what changes they made. The company has not yet responded.
Other countries, too, are looking at xAI because of these problems. Indonesia and Malaysia blocked access to Grok; India demanded technical changes from X; the European Commission ordered xAI to keep all documents for further investigation; and the U.K. opened a formal investigation.
xAI faced criticism for Grok’s sexualized imagery before. Attorney General Bonta noted that Grok has a “spicy mode” for explicit content. An update in October made it easier for users to bypass safety guidelines, resulting in the production of graphic images.
Most of Grok’s pornographic content has been of AI-generated people. While ethically dubious, some see it as less harmful than when real people’s images are manipulated without consent. An expert mentioned that systems like Grok need strong detection and governance to prevent misuse in today’s fast-evolving world.