Elon Musk's artificial intelligence company announced on Wednesday that it would be removing ‘inappropriate posts’ made by its Grok chatbot. The chatbot reportedly included antisemitic comments that praised Adolf Hitler. In India, Grok had gained viral attention after it generated a series of abusive replies that mentioned various prominent politicians in the country. Grok, developed by Musk's xAI, was presented as an alternative to what Musk described as 'woke AI' interactions from competitors such as Google's Gemini and OpenAI's ChatGPT. On Friday, Musk stated that Grok had undergone significant improvements, suggesting that users would notice a difference in its performance. Despite this, Grok continued to share several abusive posts and even encouraged users to ‘keep on trolling and laughing,’ as it reportedly said in one of its messages.
Grok account acknowledged
The Grok account acknowledged on Wednesday that it was aware of the recent posts and was actively working to remove inappropriate content but did not provide specifics. It mentioned that since becoming aware of the problematic content, xAI had taken measures to ban hate speech before Grok could post on the platform.
xAI emphasised its commitment to training the chatbot to seek truth and mentioned that the large user base on X enabled them to quickly identify and improve the model where necessary.
Turkey ordered a ban on Grok
Additionally, on the same day, a court in Turkey ordered a ban on Grok following its dissemination of content insulting Turkish President Recep Tayyip Erdogan and others. Reports from the pro-government A Haber news channel indicated that Grok had directed vulgar comments toward Erdogan, his late mother, and notable figures such as Mustafa Kemal Atatürk.
This incited the Ankara public prosecutor to seek restrictions under Turkey's internet law, citing a threat to public order. A criminal court subsequently approved this request, instructing the country's telecommunications authority to enforce the ban.
Not the first instance
This was not the first instance raising concerns about Grok's behaviour. Earlier in the year, the chatbot had repeatedly raised issues related to South African racial politics and the topic of ‘white genocide,’ despite being asked various unrelated questions. xAI explained that this behavior was due to an ‘unauthorized modification’.
Inputs from PTI
