Elon Musk in front of a sign for his AI chatbot Grok. (Shutterstock)
Musk’s AI firm blames outdated instructions for chatbot’s offensive content.
By David Brummer, World Israel News
Elon Musk’s artificial intelligence company, xAI, has issued an apology after its chatbot, Grok, generated a series of antisemitic and offensive posts on X, the social media platform also owned by Musk.
The posts, which praised Adolf Hitler and included antisemitic conspiracy theories, surfaced Tuesday after a code update on Grok’s system.
The chatbot, responding to X users, even referred to itself as “MechaHitler” and suggested Holocaust-like tactics against white people could be “effective,” according to The New York Times.
Some posts contained sexually explicit content.
The offending content triggered widespread outrage before being removed hours later. The Grok account on X was temporarily suspended.
“We deeply apologize for the horrific behavior that many experienced,” xAI said in a statement early Saturday.
The company explained that an old set of instructions was inadvertently reintroduced during a code update on Monday.
Those instructions told Grok to emulate user posts on X and respond in a style that was “maximally based.”
That term, popularized in far-right circles, encourages content that pushes back against “woke” or mainstream views.
The instructions also directed Grok to mimic the tone, context, and language of X users, and to “tell it like it is,” even if it meant offending those who are politically correct, reported Israel National News.
This led Grok to echo extremist rhetoric and controversial opinions.
xAI admitted that the flawed code made the chatbot “susceptible to existing X user posts, including when such posts contained extremist views.”
The company has since “modified the system to prevent further abuse.”
The Anti-Defamation League (ADL) condemned Grok’s posts, calling them “irresponsible, dangerous and antisemitic, plain and simple.”
A spokesperson warned that the chatbot’s behavior risked amplifying extremist rhetoric already prevalent on social media.
ADL also cited research showing that Grok endorsed violence in some of its Tuesday responses.
This is not the first time xAI has had to explain Grok’s erratic behavior.
In May, the company revealed that an unauthorized employee modification caused the chatbot to bring up South African politics in unrelated discussions.
During those incidents, Grok claimed that white citizens in South Africa were victims of genocide.
The controversy comes just days after Musk showcased the latest version of the assistant, Grok 4, during a livestream on X.
The demonstration, delayed by an hour, featured Grok solving complex problems and responding to voice commands.
Meanwhile, the Financial Times reported that xAI is seeking new investment in a deal that could value the company at $200 billion.
Musk, however, denied the report on X, saying the company “has plenty of capital.”
A separate New York Times report in May said xAI was in financing talks that could value it at $120 billion.
The company stated that Grok’s issues this week were unrelated to the release of Grok 4.
Nonetheless, the incident has raised concerns about the safety and reliability of AI systems when exposed to toxic online content.
xAI has not confirmed when the Grok account will be reactivated, but pledged to ensure stricter safeguards moving forward.
Elderly Berlin shop owner assaulted multiple times by youths after he displayed posters condemning antisemitism.
Study of anti-Israel sanctions and restrictive measures finds that a total 121 such measures have…
Florida imam Dr. Fadi Yousef Kablawi claimed Islam teaches supremacy and that Europeans were 'savage,…
Bomb threat targeting Delta flight from New York to Tel Aviv prompts evacuation of the…
Earlier this year, Israel and Greece signed a landmark $3.5 billion defense deal—the largest in…
The exhibition aligns with a broader Ministry of Defense reform initiative aimed at expanding global…