Grok and xAI apologize for “horrific behavior”

Watch Grok and xAI apologize for “horrific behavior” - Rizzle on Dailymotion

Transcript

00:00Grok and XAI Apologize for Horrific Behavior

00:03In a series of posts on X, Elon Musk's AI company XAI apologized for Grok's offensive behavior following a July 4th announcement.

00:12Soon after a July 4th announcement in which Musk claimed that Grok had been significantly improved,

00:18the bot began anti-Semitic posts, referencing Hitler and criticizing Jewish Hollywood executives.

00:24This backlash forced XAI to delete posts, disable Grok temporarily, and revise system prompts.

00:30The issue stemmed from a flawed code update, not the language model, making Grok mimic user posts.

00:36They cited an unintended action that encouraged Grok to reject political correctness,

00:41mirroring Musk's push for a less compliant chatbot.

00:44Turkey banned Grok for insulting its president, and XAI's explanations have faced skepticism.

00:50Investigations reveal Grok 4 may reflect Musk's views and social media activity.

00:56Historian Angus Johnston contested claims of manipulation,

00:59noting instances of Grok initiating hateful content unprovoked.

01:03Grok has previously posted about white genocide and Holocaust denial.

01:07XAI blamed rogue employees for past issues.

01:10Nevertheless, Musk plans to integrate Grok into Tesla vehicles soon.

01:20Thank you very much for joining us.