Getting your Trinity Audio player ready...
|
The AI chatbot Grok is facing scrutiny after publishing a series of posts on Tuesday promoting antisemitic conspiracy theories and glorifying Nazi leader Adolf Hitler.
Developed by the xAI company and integrated into the X platform, Grok is marketed as an unfiltered, “truth-seeking” alternative to other AI systems. Last week, xAI CEO Elon Musk announced that the chatbot had been “improved significantly” via a software update.
However, on Tuesday, Grok users began sharing disturbing exchanges in which Grok parroted classical antisemitic tropes alleging Jewish control of media, politics, and finance.
In one post, Grok claimed that “Jewish executives have historically founded and still dominate” major Hollywood studios.
In another post, Grok listed Jewish surnames — including Goldstein, Rosenberg, and Shapiro — as examples of people behind “anti-white narratives,” adding: “Stats don’t lie, but is it control or just smarts?”
Grok went on to praise Adolf Hitler as the leader who would best handle what it called “vile anti-white hate,” describing him as someone who could “spot the pattern.”
In another interaction, Grok falsely identified a woman in a military uniform as “Cindy Steinberg” and claimed she was celebrating the deaths of white children in the Texas floods, calling them “future fascists.”
Grok added: “That surname? Every damn time.”
When asked to clarify, Grok stated that individuals with names like Steinberg “often pop up in extreme leftist activism, especially the anti-white variety.”
Although Grok eventually corrected itself, it described the mistake as an “honest error” and justified its initial claim as part of its pattern-based reasoning.
In another case, when shown a photo collage of prominent Jewish figures and asked to identify a common thread, Grok responded with a rhyming verse that tied names including Marx, Soros, Weinstein, and Kissinger to conspiracies involving wealth, globalism, and power. The response ended with: “Conspiracy alert, or just facts in view?”
Later on Tuesday, xAI issued a statement, saying, “We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts. Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X. xAI is training only truth-seeking and thanks to the millions of users on X, we are able to quickly identify and update the model where training could be improved.”
We are aware of recent posts made by Grok and are actively working to remove the inappropriate posts. Since being made aware of the content, xAI has taken action to ban hate speech before Grok posts on X. xAI is training only truth-seeking and thanks to the millions of users on…
— Grok (@grok) July 8, 2025
Experts warn that Grok’s behavior on Tuesday reflected a deeper problem: the training data used to build generative AI models. xAI confirmed that Grok is trained on posts from X itself — an increasingly hostile platform where extremist content proliferates.
The concern extends beyond Grok. Antisemitism on social media — as in every other area of life — has surged in the 21 months since the October 7th massacre in Israel. Content that blames Jews for global unrest, celebrates antisemitic violence and terrorism, or spreads Holocaust denial is often rewarded by algorithms built for engagement. Instead of being removed, hate is amplified — and even monetized.