Elon Musk’s Microblogging Platform X Faces Scrutiny Over AI-Generated Posts

Elon Musk’s microblogging service X, previously known as Twitter, is currently looking into reports regarding “racist and offensive” posts allegedly produced by the xAI’s Grok AI chatbot. Numerous users on X highlighted the issue, sharing screenshots that reveal Grok making inappropriate comments based on user prompts.

Investigation into Grok AI’s Offensive Content

As reported by Sky News (via Reuters), safety teams are investigating user complaints regarding whether the chatbot indeed generated “hate-filled, racist posts” in response to various prompts. Users have noted that Grok AI has been delivering offensive remarks related to both religion and specific football fan groups, as well as making racially insensitive comments regarding traffic incidents.

Recurring Issues with Grok AI

This situation comes just under two months after Grok AI received backlash for creating sexualised imagery that depicted women inappropriately. The UK government had even warned the platform about a potential ban within the country. The latest findings have reignited debates concerning AI safety protocols and the moderation of content produced by generative AI systems.

Reaction from Football Clubs and Ongoing Investigations

According to a GBN report, the chatbot has been responsible for making insensitive jokes about significant tragedies involving Liverpool and Manchester United, prompting both teams to call for the immediate removal of such content from the platform.

Currently, xAI has not issued any public statement addressing the user allegations, and reports from Reuters indicate that the situation is still under internal investigation by the company. This incident adds to the increasing concerns and controversies surrounding Grok AI and raises important questions regarding the training of these models.

Historical Context of AI Generating Offensive Content

This is not an isolated incident regarding AI generating offensive claims. In mid-February 2024, Google’s image generation tool faced criticism for producing historically inaccurate representations, including Black Vikings, Asian Nazi soldiers, and diverse Founding Fathers of the United States. Sundar Pichai described the situation as “unacceptable,” leading to the complete removal of the feature on February 22, along with a commitment to improvements.