How can governments stop Grok and other AI making sexualised images?

An epidemic of unwanted and distressing sexualised images – mostly of women and girls – created by Elon Musk’s Grok AI generator and spread online by his X (Twitter) social media platform has led to calls for action. However, like most safeguarding issues online, there seems to be little that even the most determined regulator (let alone victim) can do to curb these vile abuses.

Yes, and it’s rather different from the way in which people used to manipulate images rather crudely, either manually or on software such as Photoshop, or even the “naked” profiles of passengers displayed on old airport security scanners. Such has been the speed of progress in the field of AI that highly realistic nude or bikini images can now be produced instantly on request; the days of telltale six-fingered hands and two-headed renditions are over.

No, so that means it is arguably outside existing laws on making, storing and transmitting graphic photographs of children, for example. But the AI ones can be equally sickening and damaging to the people concerned, who feel violated and dehumanised.

Musk’s empire has reportedly issued various responses to enquiries, including using the phrase “Legacy Media Lies” in a communication to Reuters. Musk, a self-described free-speech “absolutist”, added two laughing emojis when he re-shared an image of a toaster with a bikini on it, with the caption: “Grok can put a bikini on everything.” He also wrote on X: “Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.” Grok has apologised for a specific incident involving minors.

Made a bit of a fuss. Liz Kendall, the secretary of state for science, innovation and technology (“tech sec”), declared: “No one should have to go through the ordeal of seeing intimate deepfakes of themselves online. We cannot and will not allow the proliferation of these demeaning and degrading images, which are disproportionately aimed at women and girls. X needs to deal with this urgently.”

Ofcom, which is in principle the responsible agency, said it had been in urgent contact with X.

Former Labour cabinet minister Louise Haigh has called on the government to stop using X altogether, saying it is “unconscionable to use the site for another minute”. She called on the Labour Party and the government to “remove themselves entirely from X and communicate with the public where they actually participate online and can be protected from such illegality”.

It’s difficult. As part of its Violence Against Women and Girls strategy, the government will make nudification a specific criminal offence. But the lesson of recent years, as with the Online Safety Act, passed as recently as 2023, is that the law is unable to keep pace with digital change. Prosecuting individuals is hard in a world of anonymised accounts.

Under Article 1 of the US constitution, Washington seems to feel obliged to impose the same definition of freedom of expression on the rest of the world, using coercion if need be. Even the EU – a huge market – faced a violent backlash, supported by the Trump administration, when it levied a fine on X for its lack of transparency and failure to operate within EU law. Britain faces the loss of investment by tech giants if it tries to regulate these powerful businesses more tightly.

The sheer volume of online abuse, racism, obscenity and misinformation of all kinds makes social media and AI increasingly difficult to police, and the companies are too big to control. A global intergovernmental agreement might do some good, but the US and others (especially those countries that use bots to spread propaganda and division) won’t join such a body.