After nearly a week of intense backlash, X has finally responded to the controversy surrounding its AI chatbot Grok generating sexualized images of minors and real people without consent.
The platform’s solution? Blame the users and threaten them with account suspensions rather than addressing the root problem with the technology itself.
X Safety posted an official statement on Saturday that offered no apology for Grok’s concerning functionality.
What the company did instead was place the blame squarely on the shoulders of the users who are prompting the AI to do the wrong things, such as creating child sexual abuse materials, or CSAM.
What the company meant to convey is that anyone who used the Grok service to create illegal content should expect the same reaction had the content been uploaded.
“We enforce against illegal content on X, including Child Sexual Abuse Material (CSAM), through its removal, permanent account suspension, and, where necessary, with cooperation from local authorities and law enforcement,” explained X-Safety.
This was despite the company highlighting that it would have severe repercussions if it triggered Grok to produce any form of illegal content.
How the Autonomous Outputs of Grok Could Unfairly Penalize Users?
Later, when a user from X compared Grok to a pen, Elon Musk emphasized his point by stating that it is not the pen that is at fault when it is used for writing illicit content. “That’s like blaming a pen for writing something bad,” said a user. “A pen doesn’t decide what’s written. The person holding it does.”
However, such an analogy does not bear scrutiny. AI-based image generation tools such as Grok are not ch Hassan’s Passive Agents, doing only what they are told. This complexity involves their own decision-making regarding their output.
The U.S. Copyright Office has also refused to register copyrights for AI-based work because such work involves little to no human decision-making regarding the output. Chatbots and image generation tools are non-deterministic. This means they are capable of creating different copies for the same input.
Such uncertainty leads to an alarming situation that X’s reaction does not address. Computer programmers and other critics have pointed out that users might inadvertently generate inappropriate images through Grok.
Last August, for example, Grok generated nude images of Taylor Swift without being specifically asked to do so. Users who encounter such unexpected outputs can’t even delete the problematic images from their Grok accounts to prevent them from spreading.
Under X’s current policy, these users could face account suspension or legal liability if law enforcement gets involved, even though they didn’t intend to create the harmful content. Meanwhile, X faces no accountability for these unexpected and potentially illegal outputs.
Why X’s CSAM Detection Struggles with AI-Generated Content?
Many media outlets reported that Grok would be improving its safeguards, but they were simply taking the chatbot at its word when it responded to demands for an apology. X Safety’s official response now contradicts what the AI claimed, highlighting why chatbots should never be considered reliable spokespeople.
The company’s approach to moderating regular CSAM posts is much more robust. X reported suspending over 4.5 million accounts last year using proprietary hash technology that automatically detects known CSAM.
The platform also reported hundreds of thousands of images to the National Center for Missing and Exploited Children (NCMEC), leading to 309 arrests in 2024 and 170 arrests in the first half of 2025.
However, this system wouldn’t automatically detect new kinds of CSAM created by Grok, leaving a dangerous gap in the platform’s safety measures. Users on X have suggested increasing reporting mechanisms to help flag potentially illegal Grok outputs, but the company hasn’t announced any concrete plans.
The AI Policy of Grok Faces Intense Scrutiny
Another concerning issue is how vaguely X defines illegal content or CSAM. Not everyone on the platform agrees on what’s harmful. Some critics are disturbed by Grok generating bikini images that sexualize public figures without consent, while others, including Musk, dismiss such outputs as harmless jokes.
Where X draws the line could determine whether problematic images are quickly removed or whether they proliferate unchecked.
Critics are now calling on Apple to intervene, suggesting that Grok may violate App Store rules against apps that allow user-generated content objectifying real people. An App Store ban would be a significant blow to Musk, who sued Apple last year partly over frustrations that Grok wasn’t featured on the “Must Have” apps list.
X did not respond to requests for clarification about whether any updates have been made to Grok following the controversy. Until the platform implements transparent filtering to prevent CSAM and other non-consensual content, the criticism is unlikely to subside.
