AI regulation in Kazakhstan (2026)

Law No. 230-VIII 'On Artificial Intelligence' was signed by the President on 17 November 2025 and entered into force on 18 January 2026, making Kazakhstan among the first Central Asian states to enact a dedicated, horizontal AI statute.

The law classifies AI systems into low-, medium-, and high-risk tiers based on impact on citizen safety and rights; owners and operators self-classify, with regulatory obligations (documentation, conformity assessment, formal approval) scaling with risk level.

The law bans AI systems that use subconscious-manipulation techniques, exploit human vulnerabilities to cause harm, conduct social scoring, determine emotions without consent, or classify individuals biometrically in ways that enable discrimination.

Works generated entirely by AI without meaningful human creative input do not qualify for copyright protection; however, where a human contributes creatively (e.g., through prompt design), protection may apply. Training AI on copyrighted works is not automatically permitted and requires the absence of machine-readable prohibition from rights-holders.

The Government adopted a national AI Development Concept for 2024–2029 in July 2024, targeting a fivefold increase in AI-utilising products, training five million citizens in AI skills, and deploying sovereign LLMs (KazLLM, AlemLLM) on a national supercomputer with 2 EFlops capacity.

The Digital Code of Kazakhstan, adopted January 2026 and effective 11 July 2026, complements the AI Law by providing a unified framework for digital rights, data governance, and algorithmic decision-making, codifying principles of technological neutrality and human-rights protection.