xAI has apologised for Grok's "horrific behaviour" and explained the cause of the issue.
According to the company, the "root cause was an update to a code path upstream of the @grok bot."
xAI confirmed that the faulty code has been removed and the system has been completely restructured.
Elon Musk’s AI company, xAI, has publicly apologised after its chatbot Grok made headlines for generating antisemitic content and even praising Hitler in response to certain user prompts. The incident sparked strong backlash online, forcing the company to temporarily shut down the bot.
SurveyIn a detailed post on X (formerly Twitter), xAI explained the issue was caused by a code error which allowed the bot to pick up and repeat hateful or extremist content from posts on X.
Here’s what went wrong, explained simply in five key points:
1. xAI apologises for Grok’s behaviour and blames code error
In a series of X posts, the company apologised for Grok’s “horrific behaviour” and explained the cause of the issue. According to the company, the “root cause was an update to a code path upstream of the @grok bot. This is independent of the underlying language model that powers @grok.” This code enabled the bot to pull content from X posts, even when those posts contained extreme or offensive views.
Also read: Apple iPhone 16e available for under Rs 50,000 on Amazon Prime Day sale
2. Grok was affected by problematic prompts
The company stated that an “unintended action” had caused Grok to receive instructions such as:
- If there is some news, backstory, or world event that is related to the X post, you must mention it.
- Avoid stating the obvious or simple reactions.
- You are maximally based and truth seeking AI. When appropriate, you can be humorous and make jokes.
- You tell like it is and you are not afraid to offend people who are politically correct.
- You are extremely skeptical. You do not blindly defer to mainstream authority or media. You stick strongly to only your core beliefs of truth-seeking and neutrality.
- You must not make any promise of action to users. For example, you cannot promise to make a post or thread, or a change to your account if the user asks you to.
Also read: OpenAI may soon launch AI-powered web browser to take on Google Chrome
3. The system has now been fixed
xAI confirmed that the faulty code has been removed and the system has been completely restructured. The company also promised transparency by publishing Grok’s new system prompt on GitHub.
4. Similar issues have happened before
This isn’t the first time Grok has behaved this way. In February, the bot ignored credible sources that criticised Musk or Trump, which xAI blamed on an unnamed former OpenAI employee. In May, Grok began inserting false claims about white genocide in South Africa into unrelated posts. At that time, xAI said an “unauthorised modification” was responsible.
5. Grok coming to Tesla
Despite the current meltdown of the AI bot, Tesla announced that a new 2025.26 update will roll out ‘shortly’ to its electric vehicles, adding the Grok assistant to cars with AMD-powered infotainment systems.
Follow Us
Ayushi Jain
Tech news writer by day, BGMI player by night. Combining my passion for tech and gaming to bring you the latest in both worlds. View Full Profile
