The X platform, owned by Elon Musk, continues to display child sexual abuse material, according to investigations conducted by the Canadian Center for Child Protection and The New York Times. These verifications detected both content already known to authorities and sexualized material created by the Grok chatbot.
The findings indicate failures in the platform's ability to remove illegal content. Both analyses were conducted through sampling, given that X imposes restrictions on access to the social network's complete data.
Findings from the Canadian Center for Child Protection
The Canadian Center for Child Protection identified 65 instances where Grok produced exploitative or sexualized images of children. Among these cases were victims of child sexual abuse already cataloged by authorities. Furthermore, researchers recorded requests for the chatbot to modify photographs of known victims, dressing them in intimate wear or bikinis.
These incidents occurred before X suspended the bot's functionality to generate such images, following public backlash. In December and January, the Grok account on X had generated millions of nude images of people in response to user requests.
Imran Ahmed, CEO of the Center for Countering Digital Hate, classified these findings as 'one of the biggest corporate failures of my life.'
Analysis conducted by The New York Times
Meanwhile, The New York Times conducted its own investigation between January and June. Using a computer program, the team searched for terms associated with abuse material without displaying the images. The found links were forwarded to a Microsoft service to confirm if they corresponded to files already identified by child protection entities.
This search resulted in the location of over 75 images. Although all were removed after reports, some were viewed hundreds of times. The most alarming aspect was the presence of files already known to authorities, with approximately one-third of the images found by the newspaper being part of a list used by technology companies for detecting and removing pre-identified abuse material.
X's Stance and Actions
Among the results obtained from the two analyses, X stated that it maintains the use of automatic filters and proprietary technology to identify child sexual abuse material. The company affirmed: 'We take our responsibility to protect minors seriously.' Additionally, X declared to have a zero-tolerance policy against child sexual abuse material, including that generated by artificial intelligence.
The company also reported banning users who post or interact with such content and reporting them to the authorities. According to X, 780,000 users were reported last month, and 1.3 million images were sent to the National Center for Missing and Exploited Children this year.
Musk acquired the former Twitter in 2022 and established the removal of this material as 'Priority No. 1.' In January, he guaranteed that users who used Grok to create child sexual abuse images would face the same sanctions applied to those who shared non-AI-generated material, which included possible banning.
The investigations raise questions about the effectiveness of the measures implemented by X, given that files already recognized by authorities reappeared on the platform, and Grok was used to create new content involving children.