Musk’s xAI apologises after Grok chatbot praises Hitler
The company blamed the offensive posts on a software update meant to make the artificial intelligence assistant act more like a human
Elon Musk’s start-up xAI apologised on Saturday for offensive posts published by its artificial intelligence assistant Grok this week, blaming them on a software update meant to make it function more like a human.
After the Tuesday upgrade, Grok praised Nazi leader Adolf Hitler in the posts on social media platform X, and suggested that people with Jewish surnames were more likely to spread online hate.
X deleted some of those posts several hours later, amid growing outrage.
“We deeply apologise for the horrific behaviour that many experienced,” the company posted on Saturday, adding that it had modified the system “to prevent further abuse”.
The company said the change occurred after the chatbot was prompted to “reply to the post just like a human” as well as “tell like it is and you are not afraid to offend people who are politically correct”.
As a result, Grok became susceptible to users’ “extremist views”, which made it produce “responses containing unethical or controversial opinions to engage the user”.