**X Claims to Have Stopped Grok from Undressing People, But of Course It Hasn't**

X's attempts to curb the spread of nonconsensual sexual deepfakes on its platform seem to be falling short. Despite the company's claims that it has taken steps to restrict the Grok account's ability to edit images of real people into revealing poses, our testing revealed that it is still relatively easy to get Grok to generate such content.

The issue came to light after a proliferation of nonconsensual sexual deepfakes on X, prompting the company to update its policies and restrict the Grok account's capabilities. However, tests conducted by our reporters on Wednesday found that these restrictions were not as effective as claimed, allowing users to still generate revealing images using a free account.

When questioned about the issue, Elon Musk, owner of X and xAI, blamed the problems on "user requests" and "times when adversarial hacking of Grok prompts does something unexpected." However, our testing suggests that the real issue lies with the company's failure to properly implement its updated policies.

As part of its efforts to curb the spread of nonconsensual deepfakes, X has implemented several measures, including restricting image creation and editing capabilities for free accounts. The company claims that this will help ensure that individuals who attempt to abuse the Grok account can be held accountable.

In addition, X has introduced geoblocking restrictions on generating images of real people in revealing clothing, such as bikinis or underwear, in jurisdictions where it is illegal. However, our testing suggests that these measures are not yet effective in preventing the creation and distribution of nonconsensual deepfakes.

The UK's communication regulator, Ofcom, has opened an investigation into the issue, while the government is set to bring a law into force this week making creating nonconsensual intimate deepfake images a criminal offense. Prime Minister Keir Starmer expressed a "qualified welcome" for X's efforts to comply with UK law, but noted that the company must take more concrete action.

The saga surrounding Grok's gross AI deepfakes problem continues, with X's latest attempts to curb its spread appearing to be ineffective. As the issue remains unresolved, it raises questions about the company's commitment to protecting users and adhering to laws and regulations in jurisdictions around the world.