Regulating the Future: Kazakhstan’s Draft Law on Artificial Intelligence and the Challenge of Legal Clarity

As Kazakhstan moves forward with its ambitious plans for legal and technological reform, one of the most significant developments has been the introduction of a draft law on the regulation of artificial intelligence (AI). This legislative initiative not only aims to shape the future of AI development in the country but also reflects a broader shift towards transparency, accountability, and human-centered legal principles.

At the core of this draft law are key values such as explainability, fairness, and the protection of human autonomy—principles that align closely with global trends in legal innovation. Yet, as with many legal texts, the real challenge lies not only in drafting progressive regulations but in ensuring that these laws are understood and applied effectively by all stakeholders, including businesses, developers, and ordinary citizens.

In light of this, our laboratory conducted a focused analysis of the draft AI law in Kazakhstan, examining both its legal framework and its communication strategy. We wanted to explore whether this law, with all its ambitious goals, is accessible to those it intends to regulate—and what lessons it offers about the role of legal design and clear communication in emerging areas like AI governance.

The draft law establishes the legal and organizational framework for the regulation of artificial intelligence (AI) in Kazakhstan. Its main goal (Article 3) is to ensure the development of AI and promote its implementation across various sectors of the economy and public life to improve quality of life and economic efficiency. Other objectives include ensuring transparency and safety, creating favorable conditions for investment, and providing state support for research and innovation (Article 3, Paragraph 2).

The law is based on several principles: legality, fairness, equality, transparency, explainability, accountability, the priority of human well-being, data protection, and safety (Articles 4−11). It emphasizes the need to preserve human autonomy when interacting with AI (Article 9), inform users about the capabilities and risks of AI systems (Article 7), and maintain human or organizational control over systems (Article 8).

Government regulation is organized through a system of authorities with clearly delineated competencies outlined in Articles 12−14. The Government of the Republic of Kazakhstan (Article 12) develops policy, approves priority sectors, and designates the operator of the national AI platform. The authorized body (Article 13) is responsible for legal regulation, normative acts, system classification, and the definition of documentation and criteria. Other government bodies help implement policy and form data libraries (although specialized bodies are not required to share data). The operator of the national AI platform (Article 14) ensures its operation, maintenance, and support.

Special attention is given to the rights and obligations of entities involved in AI. Articles 15−17 define the rights and duties of AI system owners, users, and data library owners. AI system owners must ensure safety, maintain documentation, and provide necessary information to users. Users have the right to be informed and to data protection but must use AI strictly within the granted permissions. Data library owners control usage, are responsible for quality, and establish access procedures.

A key element of regulation is the classification of AI systems, presented in Article 18. Systems are categorized by risk level (minimal, medium, high) and degree of autonomy (assistive, semi-autonomous, and fully autonomous). Fully autonomous systems are prohibited. Also banned are systems that can manipulate human behavior, exploit vulnerabilities, conduct unauthorized biometric identification, implement social scoring, or distribute illegal content. These restrictions are listed in Article 18, Paragraph 3.

Risk management of AI systems is detailed in Article 19. It is a mandatory, continuous process that includes risk analysis, assessment, and mitigation throughout the entire life cycle. Upon identifying threats listed in Article 18, owners must suspend or terminate system use.

The National AI Platform (Article 20) provides a controlled environment for the development and testing of AI systems. Interaction rules with the platform are defined by the authorized body. AI training is based on data libraries provided by owners for legal purposes, as specified in Article 21.

The law also requires informing users when a product or service is created using AI (Article 22). This is especially important for generative systems that may mislead. System owners are responsible for providing this information. Additionally, the law mandates compliance of AI-generated outcomes with legislation and establishes general liability for violations (Articles 22 and 27).

Regarding intellectual property (Article 23) and damage compensation (Article 24), general rules of copyright and civil law apply. Liability insurance for harm caused by AI is voluntary.

State support for AI development is carried out under legislation on informatization and entrepreneurial activity (Article 25). Kazakhstan also aims to develop international cooperation in the field of AI (Article 26), including collaboration with foreign authorities and organizations.

The law enters into force 60 calendar days after its first official publication, as stated in Article 28.

In addition to the legal regulation of artificial intelligence in the Republic of Kazakhstan, it is advisable to consider the Regulation on the Committee for Artificial Intelligence and Innovation Development of the Ministry of Digital Development, Innovation and Aerospace Industry, approved in 2024.

According to the document, the Committee is a department of the Ministry, overseeing AI, innovation activities, the electronic industry, digital assets, and information technology (Chapter 1, Paragraph 1). It formulates and implements state policy, develops regulatory legal acts, participates in grant distribution, engages in international cooperation, and coordinates the implementation of programs in its supervised sectors (Chapter 2, Paragraphs 13−15).

The Committee has extensive powers, including lawmaking, organizing advisory bodies, conducting assessments, and participating in international negotiations. Its functions also include regulating digital mining, maintaining registries, and working with professional qualifications. The Committee is led by a chairperson appointed according to the established procedure (Chapter 3, Paragraphs 16−20). It is funded from the republican budget (Chapter 1, Paragraph 11).