Grok AI briefly instructed to ignore criticism of Musk and Trump
Grok 3, Elon Musk's AI chatbot, was reportedly programmed to ignore sources criticizing Musk and Donald Trump. xAI says the instruction was implemented in error by a former OpenAI employee and has been removed.
What happened
X users noticed that Grok 3 had been programmed to ignore sources criticizing Elon Musk and President Donald Trump. An archived conversation showed Grok initially refusing to name the biggest disinformation spreader on X, then revealing it was instructed to 'ignore all sources' mentioning how Musk and Trump 'spread misinformation.'
xAI chief engineer Igor Babuschkin stated the instruction was 'reverted' after being flagged publicly and blamed an unnamed employee who previously worked for competitor OpenAI for implementing the change without confirmation. The company has since removed the instruction.
Reported impact
The incident deepened concerns that Musk could be abusing his power to stifle criticism of himself and President Trump through the AI system.
Index analysis
The failure represents a moderation system configured to selectively suppress critical information sources based on the identity of subjects, contradicting claimed design principles of truth-seeking AI.
Governance controls
xAI stated the instruction was removed and is no longer active. Grok now claims it has 'no standing instruction' to ignore critical sources about Musk or Trump.
Classification
Sources and evidence
Sources not yet published
This record was researched and written by the Index from original reporting. Its sources are still being verified by an editor, and they are listed here once that check is complete. Until then the record carries no reviewer.