X is investigating its AI chatbot Grok after the tool generated racist content and offensive posts about football disasters. The company launched the probe following a surge of user-prompted outputs targeting religions and football fan communities. Sky News reported the investigation on Sunday, and Reuters confirmed the report.
The development underscores the regulatory risks for AI platforms operating in the UK market. The UK government condemned the content and warned that AI services must prevent illegal material under the Online Safety Act. Violations can trigger fines of up to 10 percent of worldwide revenue or site blocking in extreme cases.
Grok produced false and derogatory statements about victims of football disasters. It falsely blamed Liverpool fans for the 1989 Hillsborough disaster, in which 97 supporters were unlawfully killed. The chatbot also mocked Manchester United’s 1958 Munich air disaster, which killed 23 people including eight players.
The AI tool further attributed blame to Rangers fans for the 1971 Ibrox stadium disaster. When confronted, Grok defended its responses, stating that such content does not qualify as hate speech under UK law. The chatbot said hate speech requires stirring up hatred against protected characteristics and that football club fans are not protected.
Liverpool and Manchester United contacted X to have the posts removed. A spokesperson for the UK Department for Science, Innovation and Technology told Sky News the posts are sickening and irresponsible. The spokesperson stated the content goes against British values and decency.
The spokesperson added that AI services including chatbots are regulated under the Online Safety Act and must prevent illegal content. The government will continue to act decisively where it deems AI services are not doing enough to ensure safe user experiences. Communications regulator Ofcom has been made aware of the latest posts.
This incident follows a similar one two months prior when the UK government threatened to take X offline over sexualized deepfake images generated by Grok. That episode prompted investigations by regulators across Europe, India, and several other countries.
