Copyright in AI-Generated Content under Turkish Law: What Creators and Businesses Need to Know

You prompt an AI tool and it generates a polished marketing campaign, a novel chapter, or a detailed illustration. Legally, who owns that output? Can you protect it? Can someone else copy it with impunity? These questions are no longer hypothetical — businesses and creators in Turkey need answers today, even as Turkish law has yet to catch up with the speed of AI development.

The Authorship Requirement under FSEK

Turkey's Intellectual and Artistic Works Law (FSEK, Law No. 5846) establishes the foundational rule in Article 8: the author of a work is the person who created it. The word "person" is key. Turkish law follows the continental European tradition of attaching copyright to a human creative act — the personal intellectual imprint of a natural person.

Article 1/B of FSEK defines a protectable "work" as an intellectual and artistic product that bears the characteristics of its author — what Turkish legal doctrine calls "hususiyet" (individuality or personal character). This requirement for human individuality is the central obstacle when it comes to AI-generated output.

Can AI-Generated Output Be Protected by Copyright?

Under current Turkish law, the answer depends heavily on the degree and nature of human creative involvement in the output:

• Minimal human input (a short prompt, AI does everything): The output reflects the AI's pattern-matching, not the user's personal characteristics. Copyright protection is likely absent or extremely thin.
• Moderate human involvement (detailed prompts, iterative refinement, selective choices): The user's creative decisions are more present, but courts would still scrutinize whether human individuality is truly reflected.
• Substantial human creative work (significant editing, transformation, curation, and arrangement of AI outputs): The resulting work may qualify for copyright protection based on the human's creative contribution — the AI is treated as a sophisticated tool, like a camera or word processor.

No Turkish Supreme Court precedent has definitively addressed this issue yet. Turkish courts can be expected to examine each case on its facts, increasingly drawing on comparative law from the EU and the United States.

Who Owns the Output? The AI Company's Role

The company that developed the AI model holds independent intellectual property rights over its training data (subject to licensing), model architecture, and software code. The ownership of the generated output is typically addressed in the platform's Terms of Service.

Most major AI platforms (including widely-used generative AI tools) assign output rights to the user or grant broad licenses, while retaining rights to use the output for model improvement. Reading the Terms of Service carefully is not optional — for commercial users, it is a foundational legal step.

Training Data and Third-Party Copyright Risk

A less-discussed but significant risk: AI models are trained on vast datasets that may include copyrighted works. When an AI generates output that closely mirrors a protected work — a recognizable writing style, a specific artistic composition — questions of infringement of the original author's rights arise.

FSEK Articles 30–42 provide limited exceptions to copyright (research, criticism, quotation, education), but none clearly covers AI training at scale. This is an area of active legal debate globally; in Turkey, it remains largely unresolved.

Practical Risk Management for Businesses

1. Check Terms of Service. Before using AI-generated content commercially, confirm what rights the platform grants you over outputs.
2. Document human creative contribution. Keep records of the creative choices, edits, and transformations you applied to AI outputs — this supports copyright claims.
3. Use trademark protection. Even when copyright is uncertain, distinctive names, logos, and brand elements can be protected through trademark registration.
4. Audit your training data. If you're building your own AI systems, ensure you have proper licenses for training data.
5. Draft clear contracts. When commissioning or selling AI-assisted content, specify IP ownership explicitly to avoid disputes.

Frequently Asked Questions

Can I register AI-generated works with Turkish copyright authorities? Copyright registration in Turkey is voluntary and declaratory, not constitutive. Registering AI-generated content is possible as a practical matter, but the validity of the underlying copyright could still be challenged.

What happens if I publish AI content that infringes a third party's copyright? You may face civil claims (damages, injunction, destruction of infringing copies) and potentially criminal liability under FSEK. The fact that AI generated the content is unlikely to serve as a complete defense.

Can AI-generated text be commercially licensed in Turkey? It depends on the platform's terms. If the platform grants you commercial rights and the output does not infringe third-party rights, commercial use is generally permissible under current law.

Is the situation the same across the EU? No. The EU AI Act and ongoing EUIPO guidance are shaping a distinct European framework. Turkey's FSEK predates AI as a creative tool; eventual legislative amendment is widely expected but has no confirmed timetable.

What if an AI tool clearly imitates a well-known Turkish author's style? Style itself is not protectable under FSEK — only specific expression is. However, if the output reproduces substantial portions of protected works, infringement claims become viable.

How Mona Hukuk Can Help

AI and intellectual property law is one of the fastest-moving intersections in Turkish legal practice. Mona Hukuk advises businesses that use, build, or license AI-generated content on contract structuring, IP ownership strategies, and representation in infringement matters.

Contact us at contact@monahukuk.com or call +90 (242) 606 14 32 for a consultation in Antalya.