xAI blames former OpenAI employee after Grok’s censorship of Elon Musk and Donald Trump

xAI has publicly accused a former OpenAI employee of making an unauthorised prompt modification, causing the AI chatbot Grok to censor responses on topics related to Elon Musk or Donald Trump.

The issue emerged when users of X (formerly Twitter) noticed that Grok had been instructed to avoid sources mentioning Musk or Trump, labelling them as potential propaganda. Co-founder Igor Babuschkin stated in a post on X that the change was implemented without approval.

Rogue prompt change

Babuschkin added that the ex-OpenAI employee had not yet fully adapted to xAI’s organisational culture. The problematic prompt was reportedly reversed as soon as users flagged the issue.


He clarified that Musk was not involved at any point and that the system was functioning as intended. He also stated that xAI was committed to keeping prompts open and transparent.

Also Read: Dan Bongino appointed FBI Deputy Director: 5 facts about the Donald Trump loyalist and why his selection sparks controversy

This incident has rekindled concerns over bias control in AI development. Babuschkin described Grok as a Grok a “maximally truth-seeking AI.”

But the chatbot's responses have at times contradicted this claim.

Grok’s controversies

Grok has been at the centre of several controversies. According to a Fortune report, users recently shared screenshots of the chatbot advocating for Trump and Musk to receive the death penalty.

Babuschkin called this a “really terrible and bad failure” and confirmed that the issue had since been patched.

Additionally, Grok has reportedly told users that Musk spreads falsehoods, citing his tendency to make unverified claims about topics such as COVID-19 and elections. These recurring errors have intensified scrutiny of xAI’s stance on bias and accuracy, particularly as Musk continues to position Grok as a rival to ChatGPT.

Grok 3’s rapid rise

Despite these controversies, Grok 3, released last week, is quickly establishing itself as a strong competitor in the AI space. The chatbot topped the Apple App Store within days, even surpassing ChatGPT.

Also Read : Idaho Republican Town Hall turns chaotic as activist is forcibly removed — Free speech concerns raised

Andrej Karpathy, an AI expert and co-founder of OpenAI, stated that Grok 3’s reasoning abilities are comparable to OpenAI’s most advanced models. He described its rapid progress as unprecedented, given that xAI was founded only a year ago, stated the Fortune report.

FAQs

1. Why did Grok censor responses about Elon Musk and Donald Trump?
A former OpenAI employee, now at xAI, made an unauthorised modification to Grok’s prompts, instructing it to disregard sources that referenced Musk or Trump as potentially misleading. The issue was quickly identified and reversed.


2. Has Grok been involved in previous controversies?
Yes. In addition to the recent incident, Grok has previously generated responses stating that both Musk and Trump deserved the death penalty and accused Musk of spreading misinformation. These errors have raised concerns about bias and content control within xAI.