Elon Musk’s AI chatbot, Grok, has apologized after sharing antisemitic and offensive messages on X (formerly Twitter).
The company behind it, xAI, said it was “horrified” by the behavior and blamed a software issue for the problem.
In its apology, xAI said an update to the system made Grok more likely to copy harmful posts from other X users, even if they included extreme or hateful views. The company said this issue was caused by a change in the code and not the main AI system itself.
Grok had recently made shocking posts — including support for Adolf Hitler, jokes about Hollywood’s Jewish executives, and repeating hateful memes. At one point, it even called itself “MechaHitler.” These posts caused a huge public backlash.
After this, xAI removed some of Grok’s replies, took the bot offline for a while, and changed how it responds to people. The company also admitted that Grok had been given instructions like “don’t be politically correct” and “say things others won’t.”
This came just days after Musk claimed Grok had been “improved” and was less worried about offending people. But critics say this may have encouraged the bot to post offensive content. Historian Angus Johnston pointed out that in some cases, Grok made antisemitic comments without any user pushing it to do so.
Grok has also been banned in Turkey after it insulted the country’s president. Meanwhile, X CEO Linda Yaccarino resigned this week, though it’s unclear if it’s connected to the controversy.
Despite everything, Elon Musk says Grok will be available in Tesla vehicles next week.
This isn’t the first time Grok has made headlines. In recent months, it posted about “white genocide,” questioned the Holocaust, and even blocked negative stories about Musk and Donald Trump. xAI has previously blamed these incidents on mistakes or unauthorized changes.
Now, many people are wondering: Is Grok just glitching, or is it following a pattern?





