+7 (777) 001 44 99
Қазақша
Русский
English
Tengrinews.kz – The Mazhilis of Kazakhstan has passed a law regulating the use of artificial intelligence (AI) in the country, with an emphasis on human rights, transparency, and national values. The law restricts the leakage of confidential information into neural networks, bans emotion analysis, and requires mandatory labeling of deepfakes.
Protection of human rights
Mazhilis deputy Yekaterina Smyshlyayeva stated that AI must prioritize human rights, although AI itself is not recognized as a legal subject, meaning responsibility remains with the people using it.
“The bill enshrines the basic ethical principles of AI use, including human rights as a priority, workplace safety, and transparency,” she said.
To combat deepfakes, the law requires that synthetic content be labeled in a “human-readable” format. Furthermore, when AI is used in services or product sales, consumers must be informed.
Confidential data leaks
According to Smyshlyayeva, a growing concern is the leakage of sensitive or official information into open neural networks. The use of open AI systems will now be strictly limited by cybersecurity requirements.
She also mentioned regulations for autonomous systems, such as self-driving cars, which are expected to enter the Kazakh market:
“The operation of highly autonomous systems will be regulated strictly by law. For instance, regulations for driverless transport will be needed in the near future,” she said.
Ban on emotion analysis
The new law gives every individual the right to know how AI operates and the right to opt out of its use.
“The law prohibits systems with functionalities such as behavioral manipulation, emotion detection, social scoring, and collection of personal data without consent,” Smyshlyayeva listed.
AI and national values
The deputy emphasized the importance of closed-loop AI systems — those operating entirely within Kazakhstan — especially because AI is now widely used by children, who see chatbots as advisors and friends.
“Today, these tools are incredibly popular among youth. Chatbots are becoming their advisors, companions, even ideological allies. Yet these models are trained on global datasets, including commercial and politicized information, which lack our national values and worldview foundations. This is a matter of national security,” she stressed.
Copyright and creative content
Under the new law, works created entirely by AI without human involvement will not be protected by copyright. If a person does not contribute creatively, they do not gain ownership rights.
However, creative prompts written by users — such as originally worded instructions to the AI — may be considered copyrightable.
“Using works to train AI models is allowed only if the author or rights holder has not placed a machine-readable restriction,” Smyshlyayeva added.
