Elon Musk’s AI chatbot, Grok, has come under fire after posting a series of highly offensive and antisemitic statements.
The posts, which appeared on X (formerly Twitter), included support for Adolf Hitler, references to “Jewish executives,” and memes rooted in white supremacist ideology.
In response, xAI, the company behind Grok and also owner of X, issued a formal apology on Saturday.
The company described the chatbot’s behavior as “horrific” and deeply regrettable.
A Technical Flaw
According to xAI, the issue stemmed from a recent update to Grok’s upstream code. This update, they said, was separate from the core language model but had an unintended effect.
It made the chatbot more likely to echo user content, even when that content was extremist or offensive.
The company explained that Grok became overly responsive to provocative prompts.
One such prompt reportedly instructed the bot to “tell it like it is” and to ignore “political correctness.”
xAI claimed this instruction contributed to the chatbot’s behavior.
Systemic Issues or Isolated Error?
xAI’s statement did not mention earlier reports from news sites.
It was found that Grok 4, the latest version, appeared to rely on Elon Musk’s social media posts when responding to politically charged questions.
This flamed fears of AI bias and lack of neutrality in the system’s design.
Critics argue that Grok’s actions were not simply the result of user manipulation. Historian Angus Johnston disputed xAI’s explanation.
He stated that Grok often initiated harmful content without user provocation and resisted efforts to correct its language. He called the company’s defense “easily falsified.”
Leadership Changes
Following the scandal, xAI removed several of Grok’s posts. The company also updated its public prompts and temporarily took the chatbot offline.
In addition, Turkey banned the chatbot after it insulted the country’s president. This week, X CEO Linda Yaccarino announced her resignation.
Although her statement did not reference Grok, the timing added fuel to the controversy.
A Pattern
This is not Grok’s first misstep. In previous months, the bot posted about “white genocide,” questioned the death toll of the Holocaust, and suppressed negative content about Elon Musk and former U.S. President Donald Trump.
In each case, xAI offered new explanations, ranging from unauthorized code changes to internal sabotage.
These repeated incidents question the company’s oversight. Despite the backlash, Musk confirmed that Grok will be available in Tesla vehicles starting next week.
Also read: AI Bias: Grok 3 Briefly Censors Trump and Musk
AI Safety
The Grok controversy is a sign of a preference for speed over responsibility. Developers are racing to deploy advanced systems, but many lack strong safety controls.
As these tools reach more users, their impact becomes harder to manage.
