Following concerns expressed by a number of countries, including the United Kingdom and the United States, about the in-depth forgery of pornographic AI, the Elon Mask model AI Grok will no longer be allowed to edit human images that are exposed. A few hours before the change was announced, the California Supreme Prosecutor indicated that an investigation was under way into the spread of pornographic, in-depth forgery, including children, generated by Grok.
In a statement, California Attorney General Robert Bonta quoted an analysis that found that “more than half of the 20,000 images generated by xAI between Christmas and the new year were depicted as exposed people”, some of them as minors. Platform X issued a bulletin stating: “We have implemented technical measures to prevent Grok account numbers from allowing the editing of images of real people with exposed clothing such as Bikini. This restriction applies to all users, including paid subscribers.” On Wednesday, X reiterated that only fee-paying users could use Grok to edit images on the platform and indicated that this would add an additional layer of protection by helping to ensure that those who attempted to abuse Grok were held accountable for violations of the law or X policy.
X’s statement stated that under the law of the jurisdiction in which the user was located, attempts to use Grok to create images of real people wearing bikini, underwear and similar clothing would be prohibited. Mask wrote on Wednesday that with the use of the NSFW settings, Grok would allow for “imaginative upper half-naked adults”, which is consistent with what can be seen in the US R-rated film. “This is the standard of fact in the United States. In other regions, national laws will differ.” Grok’s photo-editor function has been criticized by global regulation since it came online. Last week, Malaysia and Indonesia became the first countries to ban the Grok AI tool, after users claimed that photographs had been tampered with without consent to create open images. The British Communications Authority indicated on Monday that it would investigate whether X had failed to comply with English law in the relevant photographs. British Prime Minister Kiir Stammer warned that X might lose “self-regulatory power” in the strong opposition generated by the AI image. Mask had previously defended X, writing that critics were “just trying to suppress freedom of expression” and attaching two photos produced by AI showing Kiir Stammer wearing a bikini.
The policy researcher, Ryana Pfajivkorn, noted that she was surprised that X had taken so long to deploy the new Grok security measure, arguing that the editorial function should have been removed at the beginning of the abuse. She asked how X would implement its new policy. For example, how would the AI model know if an image was real? What will be done when users fail to comply. She also stated that Mask had failed to present the company in a serious manner.