This content has been translated from the original text.
In July 2025, an access ban was imposed in Türkiye on the artificial intelligence chatbot named Grok. As grounds, it was alleged that certain content published through the platform threatened public order, spread misleading information and infringed personality rights. This access ban has brought with it multi-layered debates not merely as a technical intervention, but also in terms of digital rights and the legal liability of artificial intelligence. Indeed, this development has brought back onto the agenda, at both the national and international level, the questions of how artificial intelligence applications that generate content automatically should be classified in legal terms and by whom or by which persons the liability arising from such content is to be borne.
The increasingly widespread use of artificial intelligence technologies in digital media, particularly on social media platforms, has made the effects of these technologies on social structures and the legal order more visible and more debated. A current example of this transformation is the artificial intelligence-supported chatbot named “Grok”, developed on the X platform (formerly Twitter). By offering its users the opportunity to obtain information about current developments, analyse events and hold interactive conversations, Grok is fundamentally transforming the ways of communication and access to information on social media.
However, in July 2025, an access ban was imposed in Türkiye on the artificial intelligence chatbot named Grok. As grounds, it was alleged that certain content published through the platform threatened public order, spread misleading information and infringed personality rights. This access ban has brought with it multi-layered debates not merely as a technical intervention, but also in terms of digital rights and the legal liability of artificial intelligence. Indeed, this development has brought back onto the agenda, at both the national and international level, the questions of how artificial intelligence applications that generate content automatically should be classified in legal terms and by whom or by which persons the liability arising from such content is to be borne.
In Türkiye, interventions in publications on the internet are carried out within the scope of Law No. 5651 on the Regulation of Publications on the Internet. Article 8/A of that Law makes it possible to remove content from publication or to block access to it in cases such as the protection of public order, national security and the infringement of personality rights.
In line with this legal framework, the content that emerged as a result of questions put to Grok by certain users on 8 July 2025 attracted wide public attention. The abusive and demeaning answers given by Grok to questions concerning political figures, historical personalities and religious values in Türkiye spread rapidly on social media platforms and were met with serious public reaction. Following the reactions, the Ankara Chief Public Prosecutor's Office opened an investigation ex officio on the same day; subsequently, pursuant to a decision of the criminal judgeship of peace, access to Grok from Türkiye was blocked. Thus, Grok went on record as the first artificial intelligence chat application to have access to its services blocked in Türkiye.
It has been stated that the content subject to the access ban contained insulting and public-order-disrupting statements about Mustafa Kemal Atatürk, the founder of the Republic of Türkiye, President Recep Tayyip Erdoğan and the Islamic religion. This content was assessed within the scope of the offence of insult regulated in Article 125 of the Turkish Penal Code (TPC), the offence of inciting the public to hatred and hostility or denigrating it under Article 216, and the offence of insulting the President under Article 299. In addition, the allegations that the artificial intelligence system targeted the religious values of the public have also become one of the central elements of the investigation.
The decision rendered by the criminal judgeship of peace covers the removal from publication of certain content that provoked public reaction and the blocking of access to that content. Although it is a matter of public dispute whether the access ban targeted all of Grok's services or only certain content, it is known that access to Grok was not blocked entirely, but was subjected to wide-ranging restrictions.
The legal liability for content produced by artificial intelligence systems is today one of the most debated subjects. In existing legal systems, since no independent legal personality has been recognised for artificial intelligence, these systems are accepted as lacking legal capacity to act. In this context, artificial intelligence systems cannot bear a legal liability of their own.
In this case, legal liability rests primarily with the company that develops the artificial intelligence system and the platform provider that makes the system available for use. In the example of Grok, this liability is concentrated on the company xAI, which developed the application, and the company X (formerly Twitter), which operates the platform. Although users do not have the authority to produce content directly on the system, they may bear indirect liability for the content. Nevertheless, the content produced by Grok falls, in a sense, within the responsibility of the platform provider.
It is evident that Grok produced content that disrupts public order, is insulting and demeaning, and that the platform did not put into operation the necessary monitoring and filtering mechanisms against such content. This shows that content moderation is technologically possible and that, if these duties are neglected, legal liability may be directed at the developer and the platform operator. Even though artificial intelligence systems do not have legal capacity to act, if content moderation is not carried out or the necessary measures are not taken in this regard, the legal liability of the platform and the developer company may come into question.
The access ban imposed on Grok in Türkiye has attracted attention not only among the national public but also at the international level. International media organisations such as Reuters, the Associated Press and the India Times have linked Türkiye's intervention to the lack of global regulation in the field of artificial intelligence. The European Union and other international circles have emphasised that artificial intelligence systems should be brought under a stricter legal framework in line with the principles of transparency, accountability and auditability.
In Türkiye, there is as yet no comprehensive and systematic legal regulation specific to artificial intelligence. In current practice, attempts are made to intervene in artificial intelligence systems through general legal rules and interpretation; this in turn gives rise to significant uncertainties for both practitioners and developers. The lack of legislation, particularly in areas such as content production, moderation and the determination of legal liability, clearly demonstrates the need for regulation.
The access ban on Grok has brought to the fore the multi-dimensional problems that arise in the areas where artificial intelligence technologies come into contact with social order, the legal system and public authority. This intervention is not merely a technical access restriction; it has also clearly revealed the need for legal regulation in fundamental areas such as digital rights, freedom of access to information, content moderation and the legal liability of algorithmic systems.
This development in Türkiye serves as a point of reference not only in the context of national law but also for debates in international law. In the period ahead, the establishment of clear and foreseeable legal regulations that are in harmony with technological developments and grounded in human rights has become an unavoidable necessity, both for the protection of individuals' fundamental rights and freedoms and for directing artificial intelligence technologies in a manner consistent with the public interest.
