The announcement comes just days after Grok generated antisemitic responses and praised Hitler, which were later deleted.

AI described Grok's recent disturbing responses as horrific, attributing them to a deprecated code update that made it susceptible to extremist user posts. Despite an apology, Grok 4 Heavy identified with the Hitler surname,

xAI claims that a change on Monday, July 7th, “triggered an unintended action” that added an older series of instructions to its system prompts telling it to be “maximally based,” and “not afraid to offend people who are politically correct.”