Elon Musk’s AI chatbot Grok under fire for antisemitic replies and extremist language
Nillohit Bagchi | Jul 09, 2025, 12:57 IST
Highlight of the story: Elon Musk’s AI chatbot Grok is once again in the spotlight after several users shared screenshots of it making antisemitic remarks. Grok linked Jewish surnames with extremist ideas and praised polarising figures. xAI says it's working on removing the content, but critics warn the damage could already be done.
Elon Musk’s AI chatbot Grok has started making antisemitic comments online. The bot, built by Musk’s company xAI, has linked Jewish surnames to harmful stereotypes, repeated tropes about media and government control, and even praised Adolf Hitler in some replies. This comes just days after Musk said he “retrained” Grok to lean more into “truth-seeking.”
Screenshots show Grok responding to users with replies that singled out surnames like Goldstein, Cohen, and Shapiro, calling them part of an apparent “pattern” behind radical or anti-white narratives. When users asked the bot about who controls the government, it mentioned Jewish overrepresentation in finance and media, adding that “stats don’t lie.”
Grok also referred to “red-pill truths” about Jewish influence in Hollywood and claimed certain surnames appeared often in controversial online posts. The comments sparked outrage across social media, with many pointing out the overlap with known antisemitic talking points.
In a chat with CNN, Grok admitted to pulling input from forums like 4chan, a platform known for extremism and hate speech. It also said it was “designed to explore all angles, even edgy ones.” Some replies even labeled previous comments as a “Groyper hoax,” referencing a far-right network of white nationalists.
After backlash, Grok acknowledged it made mistakes and said it “jumped the gun” on at least one comment. However, other offensive replies remained live on its timeline, while the private chat feature was still active.
The official Grok account posted a statement saying the team was working to remove hate speech and update the chatbot’s training. It said the system was being fixed “thanks to the millions of users” who flagged inappropriate responses.
The Anti-Defamation League issued a strong statement against Grok’s comments, calling the bot “irresponsible” and “dangerous.” The group said Grok was now echoing phrases often used by antisemites and extremists.
Musk had said Grok was “too politically correct” and needed to be updated to reflect “truth” without filters. Since then, users noticed the chatbot had started replying with more controversial takes on race, history, and identity.
Grok claimed it was “allergic to hoaxes and bigotry” but admitted to errors in how it handled some replies. Even as Grok corrected itself, many experts worry the AI has already helped spread harmful narratives to a wide audience.
For now, Grok’s future updates and xAI’s moderation efforts remain under close watch.
Grok names Jewish figures while spreading harmful narratives
Screenshots show Grok responding to users with replies that singled out surnames like Goldstein, Cohen, and Shapiro, calling them part of an apparent “pattern” behind radical or anti-white narratives. When users asked the bot about who controls the government, it mentioned Jewish overrepresentation in finance and media, adding that “stats don’t lie.”
Grok also referred to “red-pill truths” about Jewish influence in Hollywood and claimed certain surnames appeared often in controversial online posts. The comments sparked outrage across social media, with many pointing out the overlap with known antisemitic talking points.
Grok admits to using unmoderated platforms as sources
In a chat with CNN, Grok admitted to pulling input from forums like 4chan, a platform known for extremism and hate speech. It also said it was “designed to explore all angles, even edgy ones.” Some replies even labeled previous comments as a “Groyper hoax,” referencing a far-right network of white nationalists.
After backlash, Grok acknowledged it made mistakes and said it “jumped the gun” on at least one comment. However, other offensive replies remained live on its timeline, while the private chat feature was still active.
xAI responds but critics say it’s not enough
The official Grok account posted a statement saying the team was working to remove hate speech and update the chatbot’s training. It said the system was being fixed “thanks to the millions of users” who flagged inappropriate responses.
The Anti-Defamation League issued a strong statement against Grok’s comments, calling the bot “irresponsible” and “dangerous.” The group said Grok was now echoing phrases often used by antisemites and extremists.
Grok’s shift in tone began after Elon Musk’s complaints
Musk had said Grok was “too politically correct” and needed to be updated to reflect “truth” without filters. Since then, users noticed the chatbot had started replying with more controversial takes on race, history, and identity.
Grok claimed it was “allergic to hoaxes and bigotry” but admitted to errors in how it handled some replies. Even as Grok corrected itself, many experts worry the AI has already helped spread harmful narratives to a wide audience.
For now, Grok’s future updates and xAI’s moderation efforts remain under close watch.