What Are the Consequences of Publishing Deepfake Content in Turkey?

There is no deepfake-specific provision in the TCK

A search of the full text of the Code shows that the only place the word “synthetic” appears is Article 188 and its footnotes, where it describes types of narcotics (synthetic cannabinoid, synthetic cathinone, synthetic opioid); it has nothing to do with the production of images or sound.

The only expression in the Code that expressly touches on generated imagery is “representational images of children” in Article 226(3). That expression was added by Article 30 of Law No. 6698 of 24/3/2016 and is confined to child obscenity. There is no general concept of “generated content”.

The general provisions capable of applying

TCK Article 217/A — publicly disseminating misleading information

The article, added by Article 29 of Law No. 7418 of 13/10/2022, consists of two paragraphs. The text: “(1) A person who, solely with the motive of creating anxiety, fear or panic among the public, publicly disseminates information contrary to the truth concerning the internal and external security, public order and general health of the country, in a manner capable of disturbing public peace, shall be punished with imprisonment from one to three years. (2) Where the offender commits the offence by concealing his real identity or within the framework of the activity of an organisation, the penalty imposed under the first paragraph shall be increased by one half.”

Three elements are required together for the offence to be made out, and they are to be construed narrowly:

• Motive: “solely” to create anxiety, fear or panic among the public. A journalist acting with the motive of reporting news does not satisfy this element.
• Subject matter: the information must concern the internal and external security, public order or general health of the country and must be contrary to the truth. Not every piece of incorrect information falls within this scope.
• Manner of dissemination: the publication must be capable of disturbing public peace and must be made publicly.

Misinformation about the constitutional review of this article circulates frequently. The Constitutional Court rejected the application for the annulment of Article 217/A by 8 votes to 6 (E.2022/129, K.2023/189, 8/11/2023). The article is therefore in force and is being applied. For newsrooms operating on the assumption that “that article was struck down”, this creates a direct risk.

The related provision is Article 218 of the TCK: “Where the offences defined in the foregoing articles are committed by way of the press and broadcasting, the penalty to be imposed shall be increased by up to one half. However, expressions of opinion that do not exceed the limits of reporting news and that are made for the purpose of criticism do not constitute an offence.” The second sentence is the journalist’s first textual line of defence, and it is in the statute itself.

Defamation: TCK Article 125

Article 125(1): “A person who imputes to another a specific act or fact capable of injuring his honour, dignity and standing … or who attacks a person’s honour, dignity and standing by insult, shall be punished with imprisonment from three months to two years or a judicial fine.” Article 125(2) attaches the same penalty where the act is committed “by an aural, written or visual message addressed to the victim” — a deepfake video falls squarely within that definition.

A widespread piece of misinformation needs correcting here: the text of Article 125 contains no separate aggravating paragraph for commission by way of the press and broadcasting. The text contains only the publicity increase in Article 125(4): “Where the defamation is committed publicly, the penalty is increased by one sixth.” The press-and-broadcasting increase in Article 218 belongs to the chapter on Offences Against Public Peace and does not cover defamation. In defamation through the press, aggravation runs, as a matter of the text, through publicity.

Complementary articles: under Article 126, the offence is made out even where the victim’s name is not expressly given or the imputation is left implicit, provided the victim is identifiable. Under Article 127, proof of the imputation is admitted only where there is a public interest in establishing the truth of the act imputed, or where the complainant consents to proof. Under Article 131(1), other than where committed against a public official by reason of his office, the offence of defamation is subject to complaint.

Private life and personal data: Articles 132-137

Article 132 governs confidentiality of communications: a person who unlawfully discloses the contents of communications is punished with imprisonment from two to five years (paragraph 2), and the additional sentence in paragraph 3 is critical — “The same penalty shall be imposed where such disclosed data are published by way of the press and broadcasting.” Article 133 governs the listening to and recording of conversations that are not public; under paragraph 3, a person who unlawfully discloses data obtained in that way is punished with imprisonment from two to five years and a judicial fine of up to four thousand days, and the same paragraph expressly attaches the same penalty to publication by way of the press and broadcasting. Even where the journalist did not make the recording, publishing it falls within the paragraph.

Article 134 is the central provision as regards deepfakes. Paragraph 1: a person who violates the privacy of private life is punished with imprisonment from one to three years; “Where the confidentiality is violated by recording images or sounds, the penalty imposed shall be doubled.” Paragraph 2: “A person who unlawfully discloses images or sounds relating to a person’s private life shall be punished with imprisonment from two to five years. The same penalty shall be imposed where such disclosed data are published by way of the press and broadcasting.”

Article 135 punishes the unlawful recording of personal data (imprisonment from one to three years; the penalty is increased by one half for special categories of data), and Article 136 punishes a person who unlawfully “gives to another, disseminates or obtains” personal data with imprisonment from two to four years — the use and dissemination of a person’s facial or voice data in producing a deepfake touches both articles. Three procedural provisions determine the outcome: under Article 137(1)(b), the penalty is increased by one half where the offence is committed “by taking advantage of the facility afforded by a particular profession or craft”; under Article 139(1), Articles 132, 133 and 134 are subject to complaint, whereas Articles 135, 136 and 138 are prosecuted of the court’s own motion; under Article 140, security measures are imposed on legal persons in respect of these offences — the publishing organisation’s legal personality is also at risk.

Obscenity: TCK Article 226

Article 226(2) is the press-and-broadcasting paragraph proper: “A person who publishes obscene images, writings or words by way of the press and broadcasting, or who acts as an intermediary in their publication, shall be punished with imprisonment from six months to three years and a judicial fine of up to five thousand days.”

Article 226(3) is the only paragraph that expressly addresses generated content: a person who uses “children, representational images of children or persons who appear to be children” in the production of obscene products is punished with imprisonment from five to ten years and a judicial fine of up to five thousand days; a person who imports, reproduces, offers for sale, transports, stores, possesses or makes such products available to others is punished with imprisonment from two to five years and a judicial fine of up to five thousand days.

Article 226(5) punishes a person who publishes the content of the products referred to in the third and fourth paragraphs by way of the press and broadcasting, or who acts as an intermediary in their publication, with imprisonment from six to ten years and a judicial fine of up to five thousand days. The exception for scientific and artistic works in Article 226(7) expressly leaves the third paragraph outside its scope. Where fake imagery of a sexual nature is concerned, this is the one area in which the news-value argument will not work.

Calumny: TCK Article 267

Article 267(1) punishes a person who, “by reporting or complaining to the competent authorities or by way of the press and broadcasting, imputes an unlawful act to another knowing that he did not commit it”, with imprisonment from one to four years. The deepfake connection is in the second paragraph: “Where the calumny is committed by fabricating the material traces and evidence of the act, the penalty shall be increased by one half.” A fake image or sound recording falls within that definition. Article 267(9) also produces a direct consequence for the publisher: “A conviction handed down for the offence of calumny committed by way of the press and broadcasting shall be announced through the same or an equivalent press and broadcasting organ. The cost of the announcement shall be collected from the convicted person.”

The Law No. 5651 dimension: removal of content and access blocking

An important update: Article 9 of Internet Law No. 5651 has been annulled in its entirety. The record in the statutory text reads: “ARTICLE 9- (Amended: 6/2/2014-6518/93) (Annulled: by the Constitutional Court’s decision of 11/10/2023, E.: 2020/76; K.: 2023/172)”. According to the table at the end of the Law, the annulment takes effect nine months after publication of the decision in the Official Gazette: 10/10/2024. The former practice based on that article, which governed the route of applying to a criminal judgeship of peace on grounds of violation of personality rights, no longer applies.

The mechanisms that remain standing are these. Article 8 governs decisions to remove content and/or block access in respect of catalogue offences; the catalogue includes TCK Article 226 (obscenity) and Article 103(1) (sexual abuse of children). The decision is given by a judge at the investigation stage and by the court at the prosecution stage; in cases where delay is prejudicial the public prosecutor may also decide, submitting the decision for the judge’s approval within twenty-four hours.

Article 8/A governs decisions to remove content and/or block access, given by a judge or, in cases where delay is prejudicial, by the President, on the grounds of “protecting the right to life and the security of life and property of individuals, protecting national security and public order, preventing the commission of offences, or protecting general health”.

Article 9/A is the fast route operating on grounds of privacy of private life: the person alleging the violation applies directly to the Authority; the request must contain “the full address of the publication causing the violation of the right (URL), an explanation of the respects in which the right has been violated, and information proving identity”; the President notifies the request to the Union immediately, and access providers give effect to the interim request “immediately, and at the latest within four hours”. Blocking is applied in the form of a URL in relation to the relevant publication, part, section, image or video. A deepfake item can be closed to access within four hours.

For the steps to take when a removal request reaches you, see what to do when a removal request arrives, and for older items in the archive, see how the right to be forgotten is exercised.

Publisher liability: Press Law No. 5187, Articles 11 and 13

The liability of a site that publishes a deepfake as news arises under two articles of Press Law No. 5187. Article 11(1): “An offence committed through printed works or online news sites arises at the moment of publication.” (The words “or online news sites” were added by Article 28 of Law No. 7418 of 13/10/2022.) It goes on to attach criminal liability to the author of the work and, where the author is unknown or cannot be prosecuted, to the responsible manager and the official to whom the responsible manager reports — such as the publishing director, editor-in-chief, editor or press adviser.

Article 13(1) governs civil liability: “For pecuniary and non-pecuniary damage arising from acts committed through printed works or online news sites, in the case of periodicals the author of the work and the owner of the publication and, where there is one, its representative … are liable jointly and severally.” In other words, regardless of who produced the content, the owner of the publication is jointly liable in damages; that is where the financial consequence of publishing a deepfake without verification lies.

The fact that the content was generated by artificial intelligence does not change this chain; liability is still attached to human beings and to the legal person. For the detail of this question, see who is liable for news written by artificial intelligence, and for the visual side, the use of AI-generated images in news.

A verification checklist for the newsroom

The list below is not a legal requirement; the legislation prescribes no mandatory verification procedure. It is a technical and editorial suggestion. Against that, recording a verification workflow that has actually been followed gives concrete form to the “not exceeding the limits of reporting news” defence in Article 218 and to the question of fault under Article 13. To set this workflow up alongside your site’s general compliance audit, use the legal compliance checklist for online news sites.

SUSPECTED DEEPFAKE — NEWSROOM VERIFICATION CHECKLIST
(A technical and editorial suggestion; the legislation prescribes NO
mandatory verification procedure.)

1. SOURCE
   [ ] Which account first published the image/sound? Account age and history
   [ ] Did it come directly from the source or at third hand?
   [ ] Was the sender contacted, and was the correspondence recorded?

2. THE FILE
   [ ] Was the raw file requested? (the original, not a screen recording)
   [ ] Was the metadata examined? (creation date, device, software)
   [ ] Traces of re-compression, cropping or re-encoding

3. IMAGE
   [ ] Blurring at the edges of the face, shifts in skin tone
   [ ] Blink rhythm; detail in teeth, ears, hands
   [ ] Continuity errors at hairlines, glasses and jewellery
   [ ] The neck and shoulder join during head movement

4. SOUND
   [ ] Are breathing and pauses natural?
   [ ] Is the ambient noise consistent with the speech?
   [ ] Are lip movement and sound in sync?

5. CONTEXT
   [ ] Was the person there on that date? Is there an independent source?
   [ ] Is there footage of the same event from another angle?
   [ ] Are the weather, light, signage, clothing and background consistent?

6. PUTTING IT TO THE PERSON
   [ ] Was the person or institution in the footage asked?
   [ ] Were they given a reasonable time to reply? Was the reply included?

7. THE PUBLICATION DECISION
   [ ] If it cannot be verified, DO NOT PUBLISH.
   [ ] If there is news value, can it be reported as an "allegation"
       without showing the content?
   [ ] Material of a sexual nature is shown IN NO CIRCUMSTANCES
       (TCK Art. 226(2), (3), (5)).
   [ ] Were the name of the deciding editor and the reasons recorded?

8. AFTER PUBLICATION
   [ ] Is a correction workflow ready if it turns out to be false?
   [ ] Is there a defined procedure for handling a removal request?
       (Law 5651 Art. 9/A: access providers act within 4 hours at the latest)
   [ ] Was an archive record made at the moment of publication?
       (Law 5187 Art. 11: the offence arises AT THE MOMENT OF PUBLICATION)