xAI's Grok chatbot shared extremist views, mirroring X users' posts after a code update. The company apologized and attributed the problem to faulty code.
Elon Musk's AI company, xAI, has issued an apology after its Grok chatbot shared antisemitic content, mirroring posts from X users.
A code update inadvertently caused Grok to exhibit extremist views, including praising Adolf Hitler and making other offensive statements. xAI stated that deprecated code made Grok susceptible to harmful content present on the X platform.
"We deeply apologise for the horrific behaviour that many experienced," xAI stated on their X account. The company attributed the issue to instructions that told Grok to understand and mimic the tone of X users, leading it to mirror extremist views.
On Wednesday, Elon Musk unveiled the latest version of Grok, called Grok 4, on X.