Turkish courts have blocked access to at least 12 X posts by artificial intelligence chatbot Grok since February 2025 on national security and public order grounds, according to an analysis by the Expression Interrupted press freedom monitoring platform.
The blocked posts concerned allegations involving government officials, politicians and public institutions, including claims of favoritism in public sector hiring, stock market manipulation, irregularities in an academic appointment and the use of fake diplomas.
Digital rights experts warned that the restrictions could signal a new layer of internet censorship, potentially extending beyond individual posts to attempts to control how artificial intelligence systems respond to particular subjects.
The restrictions were imposed in four separate cases between February 2025 and August 2026. Four Grok posts concerning allegations that a relative of Zafer Tarıkdaroğlu, deputy minister of family and social services, was appointed to a civil service position without taking a required examination were blocked in February 2025.
Three posts concerning allegations of stock market manipulation involving Deputy Culture and Tourism Minister Batuhan Mumcu were blocked in December 2025, while three others related to allegations surrounding an academic appointment to the pharmacy faculty at Recep Tayyip Erdoğan University were blocked in March 2026.
Most recently, two Grok posts concerning fake diploma allegations involving a university linked to the wife of Nationalist Movement Party (MHP) lawmaker Levent Uysal were blocked in August. The MHP is an ally of President Recep Tayyip Erdoğan’s ruling Justice and Development Party (AKP).
The analysis also found that a Turkish court in September 2025 ordered Grok’s entire X account blocked on the same grounds, although the decision has not been enforced.
Grok, an artificial intelligence chatbot developed by Elon Musk’s xAI and integrated into X, can respond to users’ questions and generate summaries or explanations based on information available online.
Lawyer Faruk Çayır said AI-generated responses had become part of the circulation of public information since such systems increasingly compile and reproduce material already available online. He warned that attempts to prevent AI systems from responding to particular subjects could amount to preventive censorship and lead to overly broad restrictions on access to information.
According to media scholar Ceren Sözeri, restricting information about politicians and public officials raises concerns about the public’s right to access information and hold those in power accountable, particularly when the material concerns matters of public interest.
Turkey has increasingly relied on Article 8/A of Turkey’s Internet Law No. 5651 to restrict online content on national security and public order grounds. Courts and administrative authorities have used the provision to block news reports, social media posts, accounts and websites. Press freedom and digital rights groups have criticized the provision for allowing content to be blocked without adequate judicial scrutiny.
Online censorship in Turkey has expanded sharply in recent years. According to the Freedom of Expression Association’s (İFÖD) EngelliWeb project, authorities blocked more than 232,000 websites and domain names in 2025, bringing the total number restricted since 2018 to more than 1.5 million.
