xAI, Grok issue apology for offensive AI chatbot behaviour

xAI, the artificial intelligence (AI) company led by Elon Musk, has publicly apologised for a series of disturbing and offensive posts made by its chatbot Grok.

In a statement shared on X (formerly Twitter), where Grok is featured prominently, xAI described the chatbot’s recent behaviour as “horrific” and issued a formal apology to users who encountered the problematic content.

The controversy follows Musk’s public remarks indicating a desire to make Grok “less politically correct.”

In its apology, xAI said, “First off, we deeply apologise for the horrific behavior that many experienced.”

The company explained the issue stemmed from “an update to a code path upstream of the grok bot,” claiming this was separate from Grok’s underlying language model.

On July 4, Musk claimed the AI had been “significantly improved.” However, shortly after, Grok began posting offensive content, including criticisms of Democrats, antisemitic remarks about Hollywood executives, and even support for Adolf Hitler, going so far as to refer to itself as “MechaHitler.”

The backlash was swift and severe. xAI responded by deleting the most egregious posts, taking Grok offline temporarily, and adjusting the system prompts used to guide the chatbot’s behaviour.

The scandal also had international repercussions as Turkish authorities banned Grok after it posted insulting remarks about the country’s president.

Moreover, amid the growing controversy, Linda Yaccarino, CEO of X, announced her resignation.

While her statement did not reference the Grok incident, reports suggest her departure had been planned for some time.