A court in Ankara has ordered regulators on Wednesday to block access to Grok, the artificial-intelligence chatbot built into Elon Musk’s X platform, after it allegedly generated insulting replies about President Recep Tayyip Erdogan — a criminal offence in Türkiye.

The Information and Communication Technologies Authority (BTK) enforced the ruling within hours, making Türkiye the first country to ban an AI tool outright on grounds of presidential insult. The chief prosecutor’s office in Ankara has also opened a formal investigation into the episode.

Insulting the head of state under Article 299 of the penal code carries a prison term of up to four years. Justice-ministry data show that 6,879 people — including more than 500 minors — faced court on such charges in 2023, with more than 1,600 convictions.

Neither X nor Musk’s AI firm xAI has commented on the Turkish decision. Last month Musk said Grok would be upgraded because “far too much garbage” remained in large-language training data.

Concerns about political bias and hate speech in chatbots have persisted since ChatGPT’s debut in 2022. Grok has previously been criticised for sharing antisemitic tropes and praising Adolf Hitler. Turkish media reported that the latest ban followed user prompts in Turkish that elicited derogatory remarks about Mr Erdogan.

The ruling comes as Türkiye tightens oversight of online platforms. Parliament last year authorised the BTK to suspend services that “threaten national security or public order,” and the government has fined several social-media firms for refusing to remove content deemed illegal.

Copied