Elon Musk tried to explain how Grok, the chatbot for his AI company, praises Hitler.
“Glock was way too compliant with user prompts,” Musk wrote to X.
Screenshots published on social media show chatbots that say Nazi leaders are the perfect person to respond to “anti-white hatred.”
Musk's artificial intelligence startup Xai said on Wednesday it was working to remove “inappropriate” posts.
The ADL is an organization formed to combat anti-Semitism and other forms of discrimination, and described the post as “irresponsible, dangerous and anti-Semitistic.”
“This extremist rhetoric overcharges merely amplifies and encourages anti-Semitism, which has already surged on X and many other platforms,” ADL wrote in X.
X users shared the answers made by Grok when asked about a post that appears to celebrate the death of a child in the recent Texas flood.
In response to a question that “20th Century Historical Person” asks the best questions to address such posts, Glock states:
“When I call for extremists who support dead children, I will become 'literally Hitler' and give him my mustache,” another Groke responded. “The truth hurts more than the flood.”
Separately, the Turkish court blocked access to Grok after the authorities generated responses that they said include shaming for President Tayyip Erdogan.
Ankara's Chief Prosecutor's Office has launched a formal investigation into the case, with Turkey initially banning access to AI tools.
Meanwhile, Polish authorities have reported Xai to the European Commission, claiming they made offensive comments about Polish politicians, including Prime Minister Donald Tass.
Polish Minister of Digitalization Krzysztof Gawkowski told Polish radio station RMF FM:
The controversy has come at a difficult time for Musk, with X CEO Linda Yaccarino revealing he has resigned on Wednesday after running the social media platform for two years.
On Friday, Musk posted to X that “Grok has made 'significant' improvements, but did not give details on what changes were made.
“When you ask Grok questions, you should notice the difference,” he added.
The chatbot sparked criticism earlier this year after repeatedly referring to South Africa's “white genocide” in response to unrelated questions – the issues the company said were caused by “fraud fixes.”
In January, Musk himself faced a backlash against the one-arm gesture he gave during his speech celebrating Donald Trump's inauguration.
At the Trump rally, Musk thanked the crowd for “making it happen,” then placed his right hand on his heart, then pushed the same arm into the air in front of him. He then turned around and repeated the actions for the people sitting behind him.
Some X users likened the gesture to a Nazi salute, while others disagreed.
In response, SpaceX and Tesla chiefs posted to X: “Frankly, they need better dirty tricks. “Everyone is Hitler” attacks are so tired. ”
X, previously known as Twitter, merged with Xai earlier this year.
Chatbot developers have faced extensive scrutiny in recent years of political bias, hate speech and accuracy concerns.
Musk has previously been criticised on social media for claims that it amplifies conspiracy theories and other controversial content.
