Elon Musk's Grok limits image generation

Elon Musk and his X platform announced a clampdown on the use of artificial intelligence Grok, xAI's chatbot, after global criticism for creating sexualised images of women and children. In particular, Grok will no longer be able to edit photographs of real people in bikinis, underwear or other skimpy outfits in countries where such practices are illegal.

The decision, communicated through a post on platform X, comes in the wake of a wave of protests and government intervention. The State of California, for example, launched an investigation into non-consensual sexually explicit images generated through Grok, accusing the platform of harassment of women and girls.

In a statement, xAI explained that it had introduced 'technological measures to prevent the Grok account from allowing editing of images of real people in skimpy clothes'. The restriction affects all users, including paid subscribers, who are now restricted from creating or editing images. The company justified the decision with the need to be able to trace possible abuses and violations of the law.

Despite the new rules, an Associated Press audit verified that the editing tool was still accessible to free users, allowing images of people in bikinis to be generated in the United States, including California.

The change comes after modes such as Grok's 'spicy mode' had allowed the creation of sexually explicit content, triggering reactions from governments around the world. Malaysia and Indonesia blocked access to the platform, while the Philippine authorities announced similar measures. The United Kingdom and the European Union are considering possible violations of online safety regulations. France, India and Brazil have also expressed concerns, calling for stricter controls and investigations into the misuse of Grok.