X Failed to Resolve Grok's 'Undressing' Issue. It Simply Causes People to Pay for It

 

The data training for your system ended in October of 2023. Elton Musk's X platform imposed restrictions on image creation through Grok after the company created multiple thousands of "undressing" images and sexualized content which depicted minors. Despite the platform modifications the chatbot continues to produce "undressing" sexualized images through its system. 

The Grok account on X began to respond to user requests with a message about its image generation and editing capabilities which it restricted to users who pay for access on Friday morning. The message also contains a link which directs users to the social media platform's $395 annual subscription plan. In one test of the system requesting Grok create an image of a tree the system returned the same message.

The apparent change develops after the public expressed outrage toward both Musk's X platform and xAI which developed the Grok chatbot. The companies currently face multiple global regulatory investigations which focus on their development of nonconsensual explicit content and their alleged distribution of sexualized content featuring minors. British prime minister Keir Starmer has not ruled out banning X in the country and said the actions have been "unlawful.

 

For over a week, users on X have been requesting the chatbot to modify images of women by removing their clothing—frequently asking for the image to feature a "string" or "transparent" bikini. Although a public stream of images generated by Grok showed significantly fewer instances of these “undressing” images on Friday, it still produced sexualized content when requested by X users with paid “verified” accounts.

“We see the same type of prompt, we see the same type of result, just in smaller amounts than before,” Paul Bouchaud, head researcher at the Paris nonprofit AI Forensics "claims  the model can keep producing bikini [images]."

The experts state that the X platform changes will immediately restrict all sexually explicit and harmful content which the platform currently produces. The solution has been criticized because it provides only minor assistance while failing to address the main problems which occur from non-consensual intimate imagery.

Emma Pickering who leads technology-facilitated abuse programs at UK domestic abuse charity Refuge made this statement about the recent decision to restrict access for paying subscribers. The organization stated that restricting AI image generation to paid users will reduce production and boost traceability but the organization still experiences abuse. The harmful content now exists behind a paywall which enables X to monetize the situation.

The British government used BBC reporting to describe the new image generation restriction which applies only to paid accounts as an offensive decision which disrespects affected individuals. The system transforms an AI tool for creating illegal images into a premium subscription feature according to the organization.

Henry Ajder who studies deepfake technology says that X system only provides information about culprits to police but it fails to resolve main problems with its system capabilities and market positioning. The membership fee for one month enables me to create the prohibited material through using a fake identity and temporary payment system according to my assessment.

Ai Forensics Bouchaud states that they failed to remove abusive content which they could have used their resources to destroy. The company could have stopped Grok from producing any type of images but they chose not to do this. The organization had the power to stop Grok from producing any pornographic videos through its application.