Elon Musk’s social media platform X started testing AI-generated news summaries powered by Grok in early April. The feature has a disclaimer that Grok “can make mistakes,” and it’s made quite a few of them.
Grok’s AI news summaries have replaced X’s trending section for paying users. It works by creating headlines and one-paragraph explanations of breaking news, using the latest tweets on X. However, Grok seems to have a hard time understanding jokes, context, and misinformation.
One of Grok’s alleged selling points is its real-time access to information on X. That capability is on full display with these AI-generated news summaries, however, it seems to have a few bugs. Musk also claims Grok to be an “anti-woke” chatbot, allegedly free of bias like other AI chatbots.
Accuracy seems to be a larger issue than bias in this case. Here are the strangest mistakes Grok has made at interpreting the news so far.