Elon Musk’s xAI has restricted its AI chatbot Grok’s picture era capabilities to paying subscribers solely, following widespread condemnation over its use to create non-consensual sexualized photos of actual girls and kids.
“Image generation and editing are currently limited to paying subscribers,” Grok introduced by way of X on Friday. The restriction means the overwhelming majority of customers can not entry the characteristic. Paying, verified subscribers with bank card particulars on file can nonetheless accomplish that, however theoretically they are often recognized extra simply if the perform is misused.
Some specialists aren’t satisfied that the brand new restrictions will repair the now widespread drawback.
“The argument that providing user details and payment methods will help identify perpetrators also isn’t convincing, given how easy it is to provide false info and use temporary payment methods,” Henry Ajder, a UK-based deepfakes knowledgeable, advised Fortune. “The logic here is also reactive: it is supposed to help identify offenders after content has been generated, but it doesn’t represent any alignment or meaningful limitations to the model itself.”
X didn’t instantly reply to Fortune’s request for remark, xAI responded with the automated message: “Legacy Media Lies.”
Over the previous week actual girls have been focused at scale with customers manipulating pictures to take away clothes, place topics in bikinis, or place them in sexually specific situations with out their consent. Some victims reported feeling violated and disturbed by the development, with many saying their stories to X went unanswered and pictures remained dwell on the platform.
Researchers mentioned the size at which Grok was producing and sharing photos was unprecedented as, in contrast to different AI bots, Grok basically has a built-in distribution system within the X platform.
Ashley St. Clair, a conservative commentator and mom of one in every of Musk’s kids, was amongst these affected by the pictures. St. Clair advised Fortune that customers have been turning photos on her X profile into specific AI-generated pictures of her, together with some she mentioned depicted her as a minor. After talking out in opposition to the pictures and elevating issues about deepfakes on minors, St Clair additionally mentioned X took away her verified, paying subscribers standing with out notifying her or refunding her for the $8 per thirty days price.
“Restricting it to the paid-only user shows that they’re going to double down on this, placing an undue burden on the victims to report to law enforcement and law enforcement to use their resources to track these people down,” Ashley St Clair mentioned of the latest restrictions. “It’s also a money grab.”
St Clair mentioned that most of the accounts concentrating on her have been already verified customers: “It’s not effective at all,” she mentioned. “This is just in anticipation of more law enforcement inquiries regarding Grok image generation.”
Regulatory stress
The transfer to restrict Grok’s capabilities comes amid mounting stress from regulators worldwide. Within the U.Ok., Prime Minister Keir Starmer has indicated he’s open to banning the platform completely, describing the content material as “disgraceful” and “disgusting.” Regulators in India, Malaysia, and France have additionally launched investigations or probes.
The European Fee on Thursday ordered X to protect all inside paperwork and knowledge associated to Grok, stepping up its investigation into the platform’s content material moderation practices after describing the unfold of nonconsensual sexually specific deepfakes as “illegal,” “appalling,” and “disgusting.”
Consultants say the brand new restrictions might not fulfill regulators’ issues: “This approach is a blunt instrument that doesn’t address the root of the problem with Grok’s alignment and likely won’t cut it with regulators,” Ajder mentioned. “Limiting functionality to paying users will not stop the generation of this content; a month’s subscription is not a robust solution.”
Within the U.S., the scenario can also be more likely to check present legal guidelines, like Part 230 of the Communications Decency Act, which shields on-line suppliers from legal responsibility for content material created by customers.
Riana Pfefferkorn of Stanford’s Institute for Human-Centered Synthetic Intelligence beforehand advised Fortune that legal responsibility surrounding AI-generated photos is murky. “We have this situation where for the first time, it is the platform itself that is at scale generating non-consensual pornography of adults and minors alike,” she mentioned. “From a liability perspective as well as a PR perspective, the CSAM laws pose the biggest potential liability risk here.”
Musk has beforehand said that “anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.” Nonetheless, it stays unclear how accounts will likely be held accountable.
