Uzbekistan takes steps to regulate AI with new legislation

According to lawmakers, the draft was developed by studying the experiences of the United States, Europe, East Asia, and CIS countries, and was coordinated with relevant ministries and agencies.

The draft also includes requirements to ensure that AI-generated information resources and AI-based information systems do not violate human rights, life, health, freedom, dignity, or other inalienable rights. It further stipulates that decisions affecting human rights and freedoms should not rely solely on conclusions drawn from AI-generated resources or systems, as stated in the Legislative Chamber’s Telegram channel.

Following Kholmukhamedova’s presentation, Deputy Ekaterina Smesova emphasized the importance of AI and the necessity of the law for society. She inquired about who would oversee compliance with the law’s requirements.

"Once this law is adopted, normative-legal acts will be developed to detail how labeling will be implemented and by whom. Naturally, individuals who create AI-generated information products and publish them in media or online will be required to label them," Kholmukhamedova responded.

Deputy Zafar Khudoyberdiev asked about the specifics of labeling AI-generated information and why the proposed amendment to the Code of Administrative Responsibility did not include penalties for failing to label. Deputy Ilhom Abdullaev responded:

"This law does not designate specific bodies to oversee the AI sector. If offenses such as defamation or mockery are committed using AI, responsible agencies will handle them under existing provisions. The only change here is the form of the offense.

In Europe, numerous regulations have been adopted to regulate AI use. In the U.S., especially after the new president took office, restrictions have been lifted to promote technological development. We aim to establish the concept of AI in the legal framework and define the basic rules and actors, rather than imposing bans, restrictions, or controls."

According to Abdullaev, labeling AI products is not akin to certifying a product. The goal is to distinguish digitally created products or artworks generated by AI and to indicate their origin. The labeling procedure will be determined by a government resolution.

Deputy Aziz Akbarov proposed amending Article 46-2 of the Code of Administrative Responsibility.

"Liability is currently limited to personal data. I proposed yesterday at the faction meeting that we include liability for actions that harm the state’s image, contradict our people’s national interests, or discredit historical figures. Including this in the draft law now would be appropriate," he said.

The draft law passed its first reading, with 130 deputies voting in favor, three remaining neutral, and no votes against.