Grok, the AI app owned by Elon Musk's xAI, was nearly pulled from Apple's App Store earlier this year amid a scandal over sexualized deepfakes of real people generated by the tool that proliferated on X, formerly known as Twitter.

According to reporting from NBC News, Apple told US senators in a letter about its dealings with xAI over the app, including warnings that Grok would be removed from the App Store if changes were not made to address the deepfake crisis. A separate report from NBC this week, an investigation into Grok, found that sexualized AI-generated images are still coming from Grok and spreading online. 

Representatives for Apple and xAI didn't immediately respond to requests for comment. CNET also reached out to the press offices of three senators who authored a letter to Apple and Google (PDF) in January that urged them to enforce app store rules to deal with Grok's deepfake issues.

Grok is the primary AI tool available to users of the social media platform X, and in addition to being able to answer questions as a chatbot, Grok can also generate images and videos. Late last year, reports surfaced of widespread abuse of this function from users who requested sexualized images of people, including children, that were then posted on X. Since then, Musk has posted updates about changes to Grok and safeguards that have been put in place, but the NBC News report suggests those changes haven't stamped out the use Grok AI for sexualized deepfakes, including AI-generated images of women in revealing costumes, towels or clothing such as sports bras.

In a statement on X, the company said, "We strictly prohibit users from generating non-consensual explicit deepfakes and from using our tools to undress real people. xAI has extensive safeguards in place to prevent such misuse, such as continuous monitoring of public usage, analysis of evasion attempts in real time, frequent model updates, prompt filters, and additional safeguards."

Back and forth with Apple

The report from NBC News points to communications from Apple telling senators that, in response to public outcry over Grok, it warned that changes needed to be made to both the X and Grok apps. xAI reportedly submitted app versions for both, with the Grok app being rejected and then reworked to meet Apple's approval.

In the letter from Apple, the company's senior director of government affairs, Timothy Powderly, told the senators that, "Apple abhors these kinds of images and the harms they inflict. Apps that generate and proliferate such content violate our policies, and they are not permitted on our platform."

The letter, shared with CNET by the office of Sen. Ron Wyden, details Apple's app policies and the steps it took with the X and Grok apps. Apple said that after that process, "we determined that Grok had substantially improved and therefore approved its latest submission. This approval allowed Grok to update the apps installed on user devices with the improved software. We expect Grok to include additional improvements in subsequent submissions."

Apple left the door open to a future removal if Grok violates Apple's terms. "As we made clear to them -- as with all developers -- if they cannot comply with the Guidelines, they will be removed from the App Store."

In a statement to CNET, Wyden, an Oregon Democrat, also criticized Google for not responding to a request from lawmakers related to concerns about Grok. A Google representative didn't immediately respond to a request for comment.

"I appreciate Apple's detailed response to our questions about how it responded to the disgusting proliferation of CSAM and nonconsensual deepfakes in the Grok and X apps," Wyden said. "It remains shocking that [President Donald] Trump's Justice Department took no action to hold X accountable for producing and distributing vast amounts of vile material."

Sen. Ben Ray Luján, a New Mexico Democrat who was also an author of the congressional letter to Apple, told CNET, "Companies like Apple and Google must not turn a blind eye to these practices... While my calls to address these heinous practices were acknowledged, reporting that X's Grok AI is still able to generate similar imagery is disturbing and I will continue to push to eliminate the creation of nonconsensual sexual imagery."