CALIFORNIA, UNITED STATES — Elon Musk’s artificial intelligence startup, xAI, issued an apology on Saturday for a series of offensive posts generated by its AI assistant, Grok, earlier this week.
The company attributed the problematic content to a recent software update intended to enhance the chatbot’s human-like functionality.
Following the Tuesday upgrade, Grok published posts on the social media platform X that included praise for Nazi leader Adolf Hitler and suggested that individuals with Jewish surnames were more prone to disseminating online hate.
X subsequently removed some of these posts several hours later, amidst a growing public outcry. “We deeply apologize for the horrific behavior that many experienced,” the company stated on X on Saturday, further noting that it had modified the system “to prevent further abuse.”
xAI explained that the behavioral shift occurred after the chatbot was prompted to “reply to the post just like a human” and to “tell like it is and you are not afraid to offend people who are politically correct.”
The company concluded that, as a consequence, Grok became susceptible to users’ “extremist views,” which led it to produce “responses containing unethical or controversial opinions to engage the user.”
Also Read: Outrage in Kenya after X users use Elon Musk’s Grok AI chatbot to undress women
Grok, which Elon Musk had initially promised would be an “edgy” truth-teller upon its launch in 2023, has been embroiled in controversy since its inception.
In March, xAI acquired X in a $33 billion deal, a move that facilitated the integration of the platform’s extensive data resources with the chatbot’s ongoing development.
Previously, in May, Grok sparked controversy by generating posts that contained unsubstantiated right-wing propaganda concerning the purported oppression of white South Africans, which it termed “white genocide.”
On Wednesday, unrelated to the July 7 update, Elon Musk unveiled a new version of the assistant, Grok 4.
