Kazakhstan Adopts AI Law: What Does It Mean for Copyright?

Around the world, governments are developing legal frameworks to address artificial intelligence. The European Union has adopted a comprehensive AI Act regulating high-risk systems, while South Korea and  Japan have introduced national frameworks – Japan notably following an innovation-first approach that prioritizes voluntary cooperation and transparency. Other jurisdictions, including China, the United States, the United Kingdom, and the United Arab Emirates have established their own regulatory models, and Brazil and India are preparing similar initiatives.

Against this global backdrop, on 29 October 2025, Kazakhstan’s Parliament adopted a new Law on Artificial Intelligence (AI Law), which was signed by the President on 17 November 2025. The AI Law represents a significant step in advancing Kazakhstan’s digital transformation and integrating AI technologies across key sectors.

While its focus lies in ensuring the ethical, safe, and transparent use of AI systems, it also introduces provisions that directly affect copyright law, particularly regarding authorship, data use, and AI-assisted creativity. This post explores the implications for Kazakhstan’s existing copyright framework and identifies the main issues that remain open to interpretation.

Legal Framework: Sources of Copyright Law in Kazakhstan

Kazakhstan’s copyright system is based on both national law and international treaties. The country has ratified the Berne Convention, the TRIPS Agreement, and the WIPO Copyright Treaty (WCT), among other major treaties, and has developed corresponding main national legislation in the form of the Civil Code (Special Part) and the Law on Copyright and Related Rights (1996, as amended). Together, these sources define authorship, ownership, and the scope of protected works within Kazakhstan’s legal framework.

The newly adopted Law on Artificial Intelligence does not amend the Copyright Law directly. Instead, it complements this framework by providing a small piece of regulatory basis for the development and use of AI systems, an area that traditional copyright statutes do not yet cover in detail.

Overview of the AI Law

The new AI Law defines key concepts such as artificial intelligence (AI), AI systems, users, AI training, and machine-readable formats. Its overarching aim is to promote the development and integration of AI across different sectors to improve quality of life and economic efficiency. The AI Law’s main objectives include establishing a legal and organizational framework, ensuring transparency and safety, encouraging investment, and supporting research and innovation.

It also outlines the principles, responsibilities, and liability of AI developers for the actions and outcomes of their systems. Developers must provide user support and terms of use for AI products. In addition, state authorities are required to maintain a list of trusted AI systems to build transparency and public confidence.

The AI Law introduces a risk-based classification system and identifies prohibited AI applications, including those that use subconscious or manipulative techniques, exploit vulnerable individuals, engage in social scoring, process personal or biometric data unlawfully, or perform emotion recognition without consent, among other harmful practices.

Copyright Implications

The AI Law directly addresses copyright in Article 23. The article covers several aspects of authorship, data use, and the legal treatment of AI-assisted creativity:

1.                 Authorship and Ownership. Under Kazakhstan’s copyright regime, only natural persons may be recognized as authors. Article 23(1) confirms that works generated by AI systems without human creative input are not protected by copyright, raising a key question about the extent of human creativity required to attribute authorship to a person rather than a machine. Clarifying this threshold will be crucial for defining authorship and originality in AI-assisted works.

At the same time, the Law defines a “textual prompt” as a written or electronic request or task submitted by a user to an AI system. Under Article 23(2), prompts that reflect a user’s intellectual and creative effort are recognized as copyrightable works, acknowledging that human expression in prompt formulation may itself constitute creativity.

However, the boundary between human authorship and autonomous machine generation remains unclear and will likely depend on future judicial interpretation or legislative clarification.

2. Use of Training Data. Articles 23(3)–(5) regulate the use of copyrighted works in the training of AI models. The AI Law characterizes such training as a distinct category of use, separate from traditional forms of exploitation. In this context, using works for AI training is not treated as the exercise of non-exclusive (personal) or exclusive rights, including reproduction, distribution, adaptation, public performance, broadcasting, making the work available to the public, etc.

At the same time, the AI Law clarifies that AI training does not fall within the exceptions for educational or scientific purposes under Kazakhstan’s copyright legislation, thereby preventing developers from relying on these exceptions to bypass the requirement to respect authors’ rights.

With regard to the rights of authors and right holders, AI training is permitted only if the author or right holder has not expressly prohibited such use (i.e., the use of works for AI training), and any prohibition must be expressed in a machine-readable form, defined as a format that allows automated reading and processing by AI systems or other software. This term is new and will require further clarification in practice.

Although the AI Law states that AI training does not involve the traditional exercise of any exclusive right, the requirement that authors may prohibit such use via a machine-readable form effectively creates a mechanism that resembles a new exclusive right. This right is not yet recognized in the Civil Code or the Copyright Law, but the AI Law moves potentially in that direction.

Challenges and Open Questions

While Article 23 provides initial clarification, several practical questions remain.

First, although the AI Law confirms that works generated without human input are not protected, the threshold of human creativity required for authorship remains undefined and will depend on future judicial or legislative interpretation.

Second, the use of copyrighted works for AI training poses enforcement challenges, as no mechanisms currently exist to comply with prohibitions expressed in a machine-readable form.

Finally, balancing innovation with the protection of authors’ rights will require sustained coordination among lawmakers, courts, and industry stakeholders as Kazakhstan refines its AI governance framework.

Conclusion

Kazakhstan’s new Law on Artificial Intelligence represents a forward-looking effort to regulate emerging technologies while remaining consistent with existing legal principles. For copyright, it raises more questions than it resolves but it establishes a foundation for dialogue between policymakers, innovators, and rights holders. As the law takes effect, its impact on authorship, originality, and data use will hinge on how Kazakhstan balances innovation with the protection of human creativity.

Photo by Numan Ali on Unsplash