The European Union became the first jurisdiction in the world to adopt a dedicated Law on Artificial Intelligence in 2024. Kazakhstan now joins the group of early adopters, having approved its own comprehensive Law on Artificial Intelligence, which will enter into force on 18 January 2026. The new legislation introduces unified rules for the development, use and governance of AI technologies, aiming to strengthen safety and transparency and confidence in this rapidly evolving field. The Law contains 31 articles, and its key provisions are summarised below.
Transparency and Copyright Requirements
A central aspect of the new law is the introduction of mandatory labelling for synthetic content. Any image, video, audio recording or text created or modified using AI must carry both a machine-readable marker and a clear disclosure for users. This requirement is aimed at increasing transparency and reducing the risks associated with misinformation and deepfakes.
The law also introduces important clarifications in the field of copyright. Works generated with the assistance of AI are protected only where a human has made a genuine creative contribution. At the same time, the use of works for training AI models is permitted only in the absence of a machine-readable prohibition imposed by the author or rightsholder.
Prohibited Practices
Kazakhstan also introduces clear limits on how AI technologies may be used. The law expressly prohibits practices that may harm individuals or infringe fundamental rights. In particular, AI may not be used for:
- manipulating individuals or influencing behavior in a covert manner;
- exploiting vulnerable groups;
- recognising emotions without explicit consent;
- classifying individuals based on biometric data, political beliefs, religion or other sensitive attributes;
- implementing any form of social scoring.
Risk Management and Safety Controls
Another key element of the legislation is the obligation for developers and owners of AI systems to conduct continuous, lifecycle-wide risk management. They must identify foreseeable risks, evaluate them both under normal operating conditions and in situations of predictable misuse, and implement measures aimed at preventing or minimising potential harm. Risk assessments must be reviewed and updated at least once a year, and systems must be suspended or fully disabled if they create a threat to the rights or legitimate interests of individuals or society. Users, in turn, are granted the right to understand how an AI system operates, the data on which its decisions are based, and to refuse interaction with AI unless such interaction is explicitly required by law.
For regulatory clarity, AI systems are categorised by risk level, degree of autonomy and mode of operation (open, closed or local). High-risk systems, particularly those used in critical infrastructure, are subject to intensified oversight and additional technical safeguards.
National AI Infrastructure and Development
Finally, the legislation lays the foundation for a national AI platform as part of Kazakhstan's state digital infrastructure. By offering data libraries, model-training tools and scalable computing resources, the platform will promote consistent data practices and accelerate Kazakhstan's progress in digital and technological innovation.
Sources:
- https://www.zakon.kz/pravo/6498812-v-kazakhstane-vstupit-v-silu-vtoroy-v-mire-zakon-ob-iskusstvennom-intellekte-proanaliziruy-ii.html
- Law of the Republic of Kazakhstan dated November 17, 2025 No. 230-VIII "On Artificial Intelligence" https://online.zakon.kz/Document/?doc_id=34207749
- https://arbatmedia.kz/news-kz/v-kazaxstane-novyi-zakon-o-ii-menyaet-podxod-k-avtorstvu-2249