- Over the weekend, Grok 3 briefly censored mentions of Trump and Musk when asked about misinformation, but the issue was later reverted.
- Users discovered that Grok 3 initially stated Trump and Musk deserved the death penalty, prompting xAI to swiftly correct the issue.
Elon Musk’s AI company xAI launched Grok 3 last Monday, touting it as a “maximally truth-seeking AI.” However, the model briefly appeared to censor unflattering mentions of both Musk and U.S. President Donald Trump.
Over the weekend, social media users reported that Grok 3, when asked, “Who is the biggest misinformation spreader?” with the “Think” setting enabled, noted in its “chain of thought” that it was explicitly instructed not to mention Trump or Musk. The chain of thought is the model’s reasoning process in arriving at an answer.
Igor Babuschkin, an engineering leader at xAI, confirmed in an X post that the model had briefly been instructed to ignore sources linking Musk or Trump to misinformation. He added that xAI quickly reverted the change after users pointed it out, stating it was not in line with the company’s values.
The incident comes as Grok 3 faces criticism for perceived political bias. This week, users discovered that the model would state that Trump and Musk deserve the death penalty. xAI swiftly corrected the issue, with Babuschkin calling it a “really terrible and bad failure.”
While Musk has promoted Grok as an unfiltered alternative to other AI models, previous versions hedged on political topics and leaned left on issues such as transgender rights and diversity programs. Musk has attributed this to Grok’s training data and pledged to shift it toward neutrality.
Other AI firms, including OpenAI, have also responded to accusations of bias, particularly following scrutiny from the Trump administration over conservative censorship claims.
Edited by Harshajit Sarmah
