New York: Elon Musk’s startup, XAI, apologized on Saturday for offensive publications published by her artificial intelligence assistant Grok this week, accusing them on a software update intended to make it work more like a human.
After Tuesday upgrade, Grok congratulated the Nazi leader Adolf Hitler in the articles on the social media platform X, and suggested that people with Jewish family names were more likely to spread hatred online.
X deleted some of these messages several hours later, in the midst of increasing indignation.
“We apologize deeply for the horrible behavior that many have known,” said society on X on Saturday, adding that it had changed the system “to prevent other abuses.”
The company said that the change had taken place after the chatbot was invited to “respond to the post as a human” as well as “say like that and you are not afraid to offend people who are politically correct”.
Consequently, Grok has become sensitive to the “extremist views” of users, which made it produced “responses containing contrary or controversial opinions to engage the user”.
Grok, which Musk has promised would be a “upset” truth after its launch in 2023, was swarmed in controversy.
In March, XAI acquired X in an agreement of 33 billion dollars which enabled the company to integrate data resources from the Chatbot development platform.
In May, Grok triggered the controversy by generating positions with unwindled right propaganda on the alleged oppression of white South Africans which he called the “white genocide”.
On Wednesday, Musk unveiled a new version of the assistant, Grok 4, which was not linked to the July 7 update.
