AI-Regulation in Kazakhstan

Kazakhstan, recognizing the potential implications of AI, has embarked on a proactive approach to regulate this technology through its 'Concept for a Legal policy until 2030'. On July 16th, Minister of Digital Development, Innovations, and Aerospace Industry, Zhaslan Madiyev, announced at a government meeting that Kazakhstan could have a new "Artificial Intelligence Law" in place by the first half of 2025.

  

Two primary issues at the forefront of AI regulation are the distribution of liability for harm caused by AI and the determination of intellectual property rights for works created with AI's involvement. Kazakhstan, like many other jurisdictions, grapples with these complex questions as it seeks to balance innovation with ethical considerations and consumer protection.

  

Regarding liability, Kazakhstan currently lacks specific regulations to determine who should be held responsible for damages caused by AI-powered systems. The government is considering developing a comprehensive digital code that may address this gap and provide a framework for regulating AI effectively. 

  

Civil liability, which entails one party's legal responsibility to compensate another for harm caused by their actions or omissions, is a central concept in this context. Establishing clear guidelines for determining liability, including the type of damage inflicted, the nature of the liability (joint or individual), and the required level of causation and evidence, is crucial for ensuring fairness and accountability.

The issue of intellectual property rights for AI-generated works is equally complex. In Kazakhstan, intellectual property is defined as a “creation of the human mind” and works such as music and literature are protected under copyright law. While Kazakh law currently defines an author as a "person," AI systems are not considered natural persons and cannot be listed as co-authors or authors of intellectual property.   

International jurisdictions have also grappled with similar questions. In Canada, a recent case involving an AI-generated image challenged the notion of AI as an author. The Canadian Internet Policy and Public Interest Clinic argued that AI systems cannot be considered authors under copyright law and that the creation of the image involved purely mechanical processes, lacking human skill or judgment. The Canadian federal court is yet to deliver its verdict on this matter.

In the United States, the copyright office has previously denied copyright registration for AI-generated works, emphasizing that copyright protection is limited to human authors. 

With regard to intellectual property rights, the European Union legislation is dominated by an approach that defines the human factor as the main one for determining the ownership of rights. Thus, the European Parliament has established that AI technology is not endowed with legal capacity and human intelligence, and therefore an AI that creates works or inventions independently will not be recognized as the author of intellectual property objects. At the same time, it is noted that if AI is used only as a tool to help the author in the creation process, then the general provisions of EU intellectual property law will apply in this case.

Thus, the European Union has attempted to differentiate the approach to works created with the help of AI (AI-assisted), and to cases where AI would create works independently (so-called AI-generated works).

In the first case, as far as we understand, the ownership of intellectual property rights will undoubtedly be recognized for the person using AI to create objects of creative work. The approach remains undisclosed in the case of creating works by artificial intelligence independently.

Despite this, in 2021, the German Federal Patent Court indicated that inventions created by AI "are not excluded from patent protection, provided that a person is indicated as the inventor in the patent application." Thus, the German court recognizes that AI can be indicated as an inventor, but only as an additional inventor, the main one must necessarily indicate the person who created the invention.

Based on the above, it can be concluded that the U.S. and European approaches to relating ownership of intellectual property rights and AI have few differences - both approaches focus on the human factor and the lack of legal standing of AI per se, but the EU is taking initiatives to further regulate the technology from a more AI-friendly stance.

​One of the primary features of artificial intelligence (AI) is its reliance on data—collection, and analysis. Among the types of data used by AI systems, personal data is often included.

There is a common belief among experts that in countries where the use and protection of personal data are minimally regulated or not regulated at all, AI systems can develop more rapidly. This is because AI requires access to certain data to achieve its objectives.

In the European Union (EU), however, the situation is quite different. The EU has implemented the General Data Protection Regulation (GDPR), which is designed to protect personal data from unlawful use. In line with the GDPR, the EU has already drafted a list of prohibited AI practices. According to the proposed Artificial Intelligence Act, the following uses of AI are expected to be banned:

  

As we can see, this list would significantly limit the applications of AI within the EU. It is worth noting that the Artificial Intelligence Act has not yet come into force and is still under review. Meanwhile, there have already been instances where existing EU laws, particularly the GDPR, have been used to penalize companies for using AI in ways that violate data protection requirements.

  

In 2022, a Hungarian bank was fined a record ​​670,000 euros for using AI systems to automatically analyse customer call recordings. The AI analysed customer voices based on key words to assess their emotional state during conversations with the bank. Subsequently, the AI created rankings and determined the priority of customers with whom the bank should contact first.

While the bank believed the primary purpose of using AI was to improve customer service quality, Hungarian authorities deemed this insufficient justification for the automated processing of data. They emphasized that the only legitimate basis for such use of personal data is the voluntary and informed consent of data subjects.

This case demonstrates that companies in the EU can be held liable for the unlawful use of AI under existing legislation.

Kazakhstan has a comprehensive Personal Data Protection Law, and as recent court cases demonstrate, companies are held accountable for violations of this law. While there are no specific regulations governing the use of AI in the context of data protection in Kazakhstan, the existing law imposes significant limitations on AI applications. For instance, if an AI system requires biometric data, which is considered personal data under Kazakh law, obtaining explicit consent is mandatory.

  

Kazakhstan, as most countries around the world needs a legislation that regulates AI and its use that addresses the legal and ethical implications of this rapidly evolving technology. While the country is still developing specific regulations, its efforts to establish a framework for liability and intellectual property rights demonstrate a commitment to balancing innovation with accountability and fairness. 

  

The future of AI regulation in Kazakhstan is contingent on the upcoming legislation. We will offer a detailed analysis of the legislation once it is officially published. In the interim, we will remain vigilant for any developments and keep you informed. Our expectation is that the legislation will comprehensively address the issues outlined in this article.