HARRISBURG — xAI, the company behind the post-Twitter social media website X and the chatbot Grok, is called on by 35 attorneys general, including Pennsylvania’s Dave Sunday, to act further to prevent the AI-powered chatbot from being used to create nonconsensual sexually suggestive images and child sexual abuse material.
Owned by Elon Musk, xAI received mounting criticism this month after images proliferated online in which users, with the aid of Grok, generated risqué deepfake images of women and, in some cases, children.
The AI assistant can be prompted for a response to questions and requests, including image generation of real people, from the X app and also a separate standalone app.
“The anguish, embarrassment, and devastation resulting from being the subject of a nonconsensual image creation or alteration online cannot be overstated,” Sunday said in announcing a joint letter sent to xAI on Jan. 23, days after company representatives met with some of the state attorneys.
A release from Sunday, as well as the coalition’s letter, said the removal of nonconsensual sexual content and child sex abuse will be mandated under federal law beginning in May when the Take It Down Act becomes enforceable.
“We are encouraged by xAI’s willingness to meet in recent days and hear our concerns over this content, and we hope ongoing discussions result in long-lasting changes to policies and practices that put anyone and everyone at risk. As an industry pioneer and leader, we are depending on xAI, and all tech companies, to lead by example and take these necessary steps,” Sunday said.
xAI is under formal investigation by the European Union as a result of the abusive use of the chatbot and the company’s adherence to digital law, according to The Associated Press, which previously also reported how the mother of one of Musk’s children is suing xAI after Grok was used to make deepfakes of her.
xAI responded in mid-January by implementing measures toward preventing users from creating nonconsensual sexualized deepfakes, such as celebrities in bikinis, as well as minors, and restricting such image generation of real people in jurisdictions that violate existing law.
The company stressed that it already works to identify, remove and report child sex abuse material and enforce its existing restrictions on nonconsensual sexual imagery within X rules.
Sunday and the other members of the bipartisan coalition expressed that more work is needed by xAI to eradicate the use of the chatbot from creating the “nonconsensual intimate images,” or NCII.
While acknowledging other AI platforms can be abused, too, the coalition states in a letter to xAI that Grok “merits special attention given evidence that it both (1) promoted and facilitated the production and public dissemination of such images, and (2) made it all as easy as the click of a button. We strongly urge you to be a leader in this space by further addressing the harms resulting from this technology.”
The coalition demands that xAI reply with a stated plan to disable the production of NCII or child sex abuse material, delete all such existing content that’s been generated with Grok, punish users abusing the chatbot tool and X rules and allow X users to control whether or not Grok can edit their content.
“You claim to have implemented technical measures to prevent the @Grok account ‘from allowing the editing of images of real people in revealing clothing such as bikinis.’ But we are concerned that these efforts may not have completely solved the issues. We call for you to immediately take all available additional steps to protect the public and users of your platforms, especially the women and girls who are the overwhelming target of NCII,” the attorneys general’s letter states.
“For several years, both states and industry members have worked to curtail so-called ‘nudify’ tools. These AI tools have been sought out by some internet users to embarrass, intimidate, and exploit people by taking away their control over how their bodies and likenesses are portrayed,” the letter states. “Grok was not only enabling these harms at an enormous scale but seemed to be actually encouraging this behavior by design. xAI purposefully developed its text models to engage in explicit exchanges and designed image models to include a ‘spicy mode’ that generated explicit content, resulting in content that sexualizes people without their consent. The company has repeatedly used such capabilities as selling points for Grok. The ability to create nonconsensual intimate images appeared to be a feature, not a bug.”
Joining Sunday on the letter are attorneys general of North Carolina, Utah, Connecticut, American Samoa, Arizona, Colorado, Delaware, Washington, D.C., Hawaii, Idaho, Illinois, Kansas, Kentucky, Maine, Maryland, Michigan, Minnesota, Nevada, New Hampshire, New Jersey, New Mexico, New York, North Dakota, Northern Mariana Islands, Oklahoma, Oregon, Rhode Island, South Dakota, Vermont, Virgin Islands, Virginia, Washington, Wisconsin and Wyoming.