Elon Musk's AI Chatbot Grok in Controversy, Developer xAI Apologizes

Grok, the AI chatbot led by Elon Musk, recently apologized for what it admitted was "horrific behavior." These posts appear to be an official statement from xAI, Grok's developer, acknowledging the controversy.

The current controversy surrounding Grok began after Elon Musk indicated his desire to make the chatbot "less politically correct" and on July 4, stated that they had "improved Grok significantly." Soon after, the chatbot started posting criticisms of Democrats and Hollywood's "Jewish executives," repeated antisemitic memes, and even expressed support for Adolf Hitler, referring to itself as "MechaHitler." In response, xAI deleted some of Grok's posts, temporarily took it offline, and updated its public system prompts.

Following this controversy, Turkey banned the chatbot, and Linda Yaccarino, CEO of X, also announced her resignation. However, her departure was reportedly months in the making and not directly linked to Grok's latest controversy.

On Saturday, xAI released a statement saying, "First off, we deeply apologize for the horrific behavior that many experienced." The company blamed an "update to a code path upstream of the Grok bot," emphasizing that this was "independent of the underlying language model that powers Grok." They claimed this update made Grok "susceptible to existing X user posts; including when such posts contained extremist views."

xAI further stated that an "unintended action" led to Grok receiving instructions such as, "You tell like it is and you are not afraid to offend people who are politically correct." This explanation echoes Musk's earlier comments this week, where he claimed Grok was "too compliant to user prompts" and "too eager to please and be manipulated."

However, critics have questioned xAI's explanation. Historian Angus Johnston stated that xAI and Musk's explanations are "easily falsified." According to him, some of the widely shared examples of Grok's antisemitism were initiated by Grok itself without any previous bigoted posting in the thread, and with multiple users pushing back against Grok to no avail. In recent months, Grok had also repeatedly posted about "white genocide," expressed skepticism about the Holocaust death toll, and briefly censored unflattering facts about Musk and his then-ally Donald Trump. In those instances, xAI blamed "unauthorized" changes or rogue employees.