Large AI models are trained on huge datasets of books, news articles, web pages, images and code. Much of this data consists of copyrighted works. Authors, newspapers, artists and image agencies have sued, arguing that their works were used without permission or payment. AI companies argue that training is lawful and that models learn from works rather than copying them.
The key legal questions
- Is use for training lawful? In the US, companies rely on the “fair use” doctrine, which protects transformative uses. How it applies to AI training is being argued in court, and different cases have reached different assessments.
- Does it matter how the data was obtained? In some cases, downloading works from pirate sources became a separate issue.
- What if a model reproduces a work verbatim? A model producing a copyrighted text or image almost exactly can strengthen infringement claims.
Prominent cases
In December 2023 The New York Times sued OpenAI and Microsoft, alleging unauthorised use of its articles. Getty Images sued Stability AI in both the US and the UK, alleging unauthorised use of its images. Many authors and artists have also brought class actions against various AI companies, some of which have ended in settlements. The outcomes are closely watched because they could directly affect the AI industry’s business model. Meanwhile, many publishers and image agencies have signed licensing deals with AI companies.
Is AI output protected by copyright?
The US Copyright Office has stated that copyright requires human creativity and that content generated entirely by AI from prompts alone is not protected, though parts with significant human contribution can be. Approaches may differ in other countries. Turkish copyright law defines a work as an intellectual or artistic product “bearing the characteristics of its author”, and the status of AI-generated output has not yet been settled there.
The European Union’s approach
In the EU, text and data mining is permitted under certain conditions, but rights holders can opt out of such use of their works in a machine-readable way. The EU AI Act also requires providers of general-purpose models to publish a summary of their training data and to have a copyright policy.
Advice for content creators
- Keep your own contribution to AI-assisted content significant, and document it.
- For commercial use, read the tool’s licence terms.
- Avoid prompts that copy other people’s works exactly.
- If you don’t want your website’s content used for training, you can limit AI crawlers with tools such as robots.txt, though not every company honours this in the same way.
Different approaches between countries
There is no single rule in the world on AI and copyright. A few examples show how different the approaches can be:
- United Kingdom: Section 9(3) of the Copyright, Designs and Patents Act 1988 treats the author of a "computer-generated work" with no human author as the person by whom the arrangements necessary for its creation were undertaken. This provision was written before the AI era, and how it applies is debated.
- Japan: Article 30-4 of the Copyright Act, added in 2018, grants a broad exception for uses such as data analysis that are not aimed at enjoying the content of works; it is considered one of the most flexible rules for model training.
- China: In November 2023 the Beijing Internet Court ruled that an AI-assisted image could be protected by copyright where the user made a distinct creative contribution through prompts and settings.
- United States: In Thaler v. Perlmutter, the appeals court affirmed in March 2025 that an image generated entirely by an AI system cannot be registered without a human author.
What is the situation in Türkiye?
In Türkiye the issue is assessed under Law No. 5846 on Intellectual and Artistic Works (FSEK). The law defines a work as an intellectual or artistic product "bearing the characteristics of its author", and the author as the person who created the work. Because these definitions were written before AI, legal scholars debate whether content produced without human contribution can be protected; the general tendency is that protection depends on a human's creative contribution.
In practice, the safest path for those producing content with AI is to document their own creative contribution (selection, arrangement, rewriting) and to read the terms of use of the tool they use. This section is for general information; consult a lawyer for specific situations.
AI tools’ features, plans and prices change often. Always check the tool’s official page for current information.
This article is for general information and is not legal advice.
Sources and further reading
- U.S. Copyright Office, “Copyright and Artificial Intelligence” reports (Part 2, January 2025; Part 3, May 2025).
- The New York Times Co. v. Microsoft Corp. and OpenAI, U.S. District Court for the Southern District of New York (December 2023).
- European Union, Directive (EU) 2019/790 on copyright in the Digital Single Market, Art. 4.
- UK Copyright, Designs and Patents Act 1988, s. 9(3) – legislation.gov.uk
- Thaler v. Perlmutter, D.C. Circuit (18 March 2025)
- Law No. 5846 on Intellectual and Artistic Works – mevzuat.gov.tr
Leave a comment
Your email address will not be published. Required fields are marked *