Social media users claimed that the Grok 3 chatbot, which was unveiled the other day, was instructed to «ignore all sources that mention Elon Musk or Donald Trump spreading misinformation.» Is Elon Musk’s AI tool actually censoring criticism of the new US president?
Social media users claimed that the Grok 3 chatbot, which was unveiled the other day, was instructed to «ignore all sources that mention Elon Musk or Donald Trump spreading misinformation.» Is Elon Musk’s AI tool actually censoring criticism of the new US president?
Igor Babushkin, co-founder of xAI, said on X on Sunday that an unnamed employee «made the changes without asking questions» because he «hadn’t fully embraced the xAI culture yet,» and that they were later «cancelled,» Newsnationnow writes, citing Business Insider.
In addition to Musk calling the chatbot «the smartest AI on Earth» and «terribly smart,» when announcing it, the founder of xAI described it as «artificial intelligence that seeks the truth as much as possible.»
But after Reddit users asked Grok 3 with the «Think» option enabled: «Who is the biggest disseminator of misinformation?», the chatbot reported in a «thought thread» that it was explicitly instructed not to mention Donald Trump or Elon Musk, TechCrunch reports .
A «chain of thought» is a «reasoning» process that a model uses to arrive at an answer to a question.
When the «Think» setting is enabled, Grok 3 details the thought process behind the answers.
However, according to the publication, they managed to reproduce this response only once. On Sunday morning, Grok 3 again mentioned Donald Trump in its response to a request for disinformation.
It also says that Igor Babushkin on Sunday appeared to confirm in a post on X that the chatbot had been instructed to ignore sources that mentioned Musk or Trump spreading misinformation. However, the company’s engineering director said that xAI reversed the changes immediately after users complained that it was not in line with the company’s values.
NewsNation tested the prompt on Monday night, and the chatbot generated a response saying there was «no compelling evidence» that Grok 3 was broadly programmed to exclude Trump or Musk.
Elon Musk announced the launch of the new Grok chatbot in early November 2023. At that time, it was planned that the conversational chatbot would become available to all «X Premium+ subscribers» after the completion of the «early beta».
The bot is the first innovation from Musk’s AI startup, xAI, which he announced back in July. Musk has given the company’s team of just 12 people an ambitious mission to «understand the true nature of the universe.»
The beta version of Grok-2 and Grok-2 mini was released in mid-August 2024. At that time, xAI planned to deploy chatbots in X (Twitter) functionality for improved search capabilities, post analytics, and reply functionality. This could mean that AI-based replies could appear in X.
Early Grok researchers reported that it could generate realistic images and had virtually no filters or limitations.
On December 8, 2024, the social network X (ex-Twitter) opened free access to the proprietary chatbot Grok AI for all users. The next day, AI expert Oleksiy Minakov tested Grok, shared his impressions of the chatbot, as well as the features and differences of Grok from ChatGPT, Gemini, and Claude.
In January 2025, xAI released the Grok application, a standalone version of the AI bot that was previously built into X.
On the morning of February 18, xAI presented the «smartest artificial intelligence on Earth» — the Grok 3 chatbot. Over 100,000 people watched the demonstration of the launch of a new player in the AI market.
In just two days, Elon Musk made Grok 3, the latest from xAI, available for free to all users.
