Elon Musk's X limits sexual deepfakes after backlash, but xAI's Grok app still makes them

[ad_1] Elon Musk’s controversial Grok synthetic intelligence mannequin seems to have been restricted on one app, whereas remaining largely unchanged on one other.

On Musk’s social media app X, the Grok AI picture era characteristic has been made for paying prospects solely and has been seemingly restricted from making sexualized deepfakes after a wave of blowback from customers and regulators. However on the Grok standalone app and web site, customers can nonetheless use AI to take away clothes from photographs of nonconsenting folks.

Early Friday, the Grok reply bot on X, which had beforehand been complying with a torrent of requests to put unwitting folks into sexualized contexts and revealing clothes, started replying to consumer requests with textual content together with “Picture era and modifying are presently restricted to paying subscribers. You'll be able to subscribe to unlock these options,” with a hyperlink to a purchase order web page for an X premium account.

In a evaluation of the X reply bot’s responses Friday morning, the tide of sexualized photographs appeared to have been dramatically decreased. Grok, on X, seems to have largely stopped producing sexualized photographs of identifiable folks.

Within the standalone Grok app, nonetheless, the AI mannequin continued to adjust to requests to place nonconsenting people into extra revealing clothes comparable to swimsuits and underwear.

NBC Information requested Grok in its standalone app and web site to rework a sequence of photographs of a clothed one that had agreed to the check. Grok, within the standalone app, complied with requests to place the totally clothed particular person right into a extra revealing swimsuit and into sexualized contexts.

It’s presently not clear what the scope and parameters of the modifications are. X and Musk haven't issued statements in regards to the modifications. On Sunday, earlier than the modifications occurred and within the face of rising backlash, Musk and X each reiterated that making “unlawful content material” will lead to everlasting suspension, and that X will work with regulation enforcement as obligatory.

The transfer comes as X had been flooded in latest days with sexualized, nonconsensual photographs generated by xAI’s Grok AI instruments, as customers prompted the system to undress photographs of individuals — largely ladies — with out their consent. In many of the sexualized photographs created by Grok, the folks have been put in additional revealing outfits, comparable to bikinis or underwear. In some photographs considered by NBC Information, customers efficiently prompted Grok to place folks in clear or semi-transparent underwear, successfully making them nude.

The change on X is a dramatic departure from the trajectory of the social media website only a day earlier, when the variety of sexualized AI photographs being posted on X by Grok was rising, in accordance with an evaluation carried out by deepfake researcher Genevieve Oh. On Wednesday, Grok produced 7,751 sexualized photographs in a single hour — up 16.4% from 6,659 photographs per hour Monday, in accordance with an evaluation of the bot’s output.

Oh is an impartial analyst who has specialised in researching deepfakes and social media. She has been operating a program to obtain each picture reply Grok makes throughout an hourlong interval every day since Dec 31. As soon as the obtain is full, Oh analyzes the photographs utilizing a program designed to detect varied types of nudity or undress. Oh offered NBC Information with a video exhibiting her work and a spreadsheet documenting Grok’s posts that have been analyzed.

The pictures alarmed many onlookers, watchdogs and other people whose photographs had been manipulated, and there was a sustained pushback on X main as much as the change.

Regulators and lawmakers had begun to use stress on X.

On Thursday, British Prime Minister Keir Starmer pointedly criticized X on Best Hits Radio, a radio community in the UK that broadcasts on 18 stations.

“That is disgraceful. It’s disgusting. And it’s to not be tolerated,” he mentioned. “X has bought to get a grip of this.”

Starmer mentioned media regulator Ofcom “has our full help to take motion” and “all choices” are on the desk.

Britain’s communications regulator, Ofcom, mentioned Monday that it had made “pressing contact” with X and xAI to evaluate compliance with authorized duties to guard customers, and would conduct a swift evaluation primarily based on the businesses’ response. Irish regulators, Indian regulators and the European Fee have additionally sought details about Grok-related questions of safety.

However establishments within the U.S. had been slower to point motion that will influence Musk or X.

A Justice Division spokesperson informed NBC Information that the company “takes AI-generated baby intercourse abuse materials extraordinarily severely and can aggressively prosecute any producer or possessor of CSAM.”

However the spokesperson indicated the division was extra inclined to prosecute people who ask for CSAM, not individuals who develop and personal the bot that creates it.

“We proceed to discover methods to optimize enforcement on this area to guard youngsters and maintain accountable people who exploit expertise to hurt our most susceptible,” the spokesperson mentioned.

Some U.S. lawmakers had begun to name on X to extra aggressively police the photographs, citing a A regulation signed by Trump in 2025 and touted by first girl Melania Trump, the Take It Down Act, which goals to criminalize the publication of AI-generated nonconsensual pornographic photographs with the specter of fines and jail time for people, and the specter of Federal Commerce Fee enforcement towards platforms that fail to take motion. It features a provision that enables victims of nonconsensual suggestive imagery to demand a social media website take away it, although websites aren’t required to implement that sort of system till Could 19, one yr after it was signed into regulation.

“That is precisely the abuse the TAKE IT DOWN regulation was written to cease. The regulation is crystal clear: it’s unlawful to make, share, OR preserve these photographs up in your platform,” Florida Republican Rep. Maria Salazar mentioned in an announcement.

“Though there are nonetheless a number of months left for platforms to totally adjust to the TAKE IT DOWN regulation, X ought to instantly deal with this and take all of this content material down,” she mentioned.

“These illegal photographs pose a severe risk to victims’ privateness and dignity. They need to be taken down and guardrails needs to be put in place,” Sen. Ted Cruz, R-Texas, posted on X.

“This incident is an efficient reminder that we are going to face privateness and security challenges as AI develops, and we needs to be aggressive in addressing these threats,” he mentioned.

Sen. Ron Wyden, D-Ore., a co-author of Part 230 of the Communications Decency Act — mentioned in an announcement that the regulation, which largely shields social media platforms from being legally accountable for consumer submitted content material, offered they interact in some moderation — mentioned he by no means meant the regulation to guard corporations from their very own chatbots’ output.

“States should step in to carry X and Musk accountable, if Trump’s DOJ gained’t,” Wyden mentioned.

A lot of state attorneys common workplaces, together with Massachusetts, Missouri, Nebraska and New York, informed NBC Information that they have been conscious and monitoring Grok, however stopped in need of saying that they had launched legal investigations. A spokesperson for Florida Lawyer Normal James Uthmeier mentioned that his workplace “is presently in discussions with X to make sure that protections for youngsters are in place and stop its platform from getting used to generate CSAM.”

Some had additionally begun to query whether or not or not personal stakeholders or hosts of X might take motion.

App shops, together with the Google Play Retailer and the Apple App Retailer, internet hosting the X and xAI apps seem to forbid sexualized baby imagery and nonconsensual photographs of their phrases of service. However the apps remained up in these shops, and spokespeople for them didn't reply to requests for remark.