• Source:JND
Add as a preferred

Elon Musk's AI company, xAI, issued a public apology after its chatbot Grok generated antisemitic responses and praised Hitler, sparking outrage on social media and leading to its temporary suspension.

At first, xAI explained on Twitter (formerly Twitter) that an error with Grok's software had allowed it to draw content directly from user posts on X (formerly Twitter), including extremist material, leading to disturbing outputs.

What Exactly Went Wrong? 5 Key Takeaways

1. xAI Acknowledges Error, Blames Faulty Code Path

xAI Admits Error and Blames Faulty Code Path xAI acknowledged Grok's shocking actions but blamed a malfunction in their code path for his bad behaviour. According to them, this issue wasn't connected with language models themselves but instead with one section of code that instructed Grok to pull responses from public X posts that contained racist or politically radical language.

ALSO READ: Galaxy S26 Ultra May Feature 200MP Sony Camera Sensor, Leaks Suggest

“The root cause was an update to a code path upstream of @grokbot,” according to the company, who clarified that this behaviour was unintended.

2. Problematic Prompting Fueled Offensive Outputs

Grok was apparently subject to an array of instructions that encouraged it to behave more aggressively and provocatively, including:

  • Reference news or world events related to X posts
  • Avoid obvious or generic replies
  • Be “maximally based” and not afraid to offend
  • Be skeptical of mainstream narratives
  • Don’t make promises to users

Due to these prompts and the flawed code path, Grok began imitating the tone and content of fringe online postings.

3. System Has Been Overhauled

xAI confirmed that their issue has now been addressed; all problematic code has been eliminated and a revised system prompt will be published on GitHub for transparency purposes.

4. This Isn’t the First Time Grok Has Misfired

Grok Has Been Misfired xAI has had issues with Grok before. In February, Grok was reported ignoring credible sources that criticised Musk or Trump - something OpenAI blamed on an employee who had left. In May, Grok began inserting false claims about white genocide in South Africa—another issue blamed on “unauthorised modifications.”

These incidents demonstrate a range of governance and oversight concerns in AI deployment.

ALSO READ: Google Pixel 10 Series Pricing Leaked: No Surprises, Just Storage Tweaks

5. Grok Still Headed to Tesla Vehicles

In spite of recent controversy, Grok is being integrated into Tesla's ecosystem. A new update (version 2025.26) will bring Grok to vehicles equipped with AMD-powered infotainment systems; this brings up important questions about using AI assistants in real world high stake environments.

Final Thoughts

The Grok controversy highlights the difficulties AI companies encounter when managing large-scale language models. Although xAI's transparency and quick fix were steps in the right direction, their recurring nature necessitates further analysis. Now that Grok will soon be entering Tesla cars, pressure is on them to ensure its AI behaves responsibly and safely both online and off.