Publications
Regulating Artificial Intelligence in Uzbekistan: How the 2025 AI Bill Was Developed and Approved
1. Background: A Strategic and Political Push Toward AI Governance
Uzbekistan’s 2025 AI bill did not emerge in a vacuum. It is the culmination of several years of strategic planning and growing political attention to digital transformation, data governance, and the ethical dimensions of new technologies.
- expanding the national AI market to USD 1.5 billion by 2030;
- ensuring that at least 10% of public services on the Unified Portal are powered by AI;
- establishing 10 AI research laboratories and deploying high-performance computing infrastructure;
- lifting Uzbekistan into the top 50 countries in the Government AI Readiness Index.
A key starting point is the Presidential Resolution No. PP-358 (14 October 2024), which approved the Strategy for the Development of Artificial Intelligence Technologies until 2030. The Strategy sets out an ambitious vision:
The document prioritizes:
- building a regulatory and ethical framework for AI technologies;
- investing in data infrastructure and computing resources;
- developing human capital and digital literacy nationwide;
- applying AI in finance, tax and customs administration, healthcare, agriculture, and energy.
In January 2025, at the Digital Almaty 2025 forum, Prime Minister Abdulla Aripov reaffirmed that Uzbekistan intended to introduce ethical standards for AI use in 2025, highlighting three pillars: infrastructure, workforce development, and support for innovative startups.
By early 2025, therefore, the country already had:
- a long-term strategic framework;
- a political mandate emphasizing ethics and rights;
- an expectation that legal boundaries for AI would be established.
This context shaped the AI legislation introduced that spring.
November 14, 2025
2. The Legislative Initiative: What the First Version of the Bill Contained
On 21 March 2025, a group of members of the Legislative Chamber of the Oliy Majlis introduced an AI regulation bill using their right of legislative initiative.
The initial draft had several notable components:
2.1. A legal definition of “artificial intelligence”
The bill established a formal definition of AI and outlined key state policy directions in the field.
2.2. Protection of human rights and dignity
The text established clear boundaries:
- AI-based systems and information resources may not harm a person’s life, health, liberty, dignity, or any other fundamental rights;
- decisions affecting individual rights cannot be made solely on the basis of AI outputs and must include human involvement.
This reflected a concern widely shared among policymakers: automated systems should support, not replace, human judgment in consequential matters.
2.3. Mandatory labelling of AI-generated content
The draft bill introduced obligatory disclosure for any content created using AI—texts, images, videos, audio—designed to prevent deception and limit the spread of deepfakes.
2.4. Personal data protection and liability
The proposal included legal responsibility for:
- unlawful processing of personal data using AI technologies;
- dissemination of such data via the media or the internet.
From the outset, the bill was general and principle-based rather than technically prescriptive.
3. Parliamentary Review in the Legislative Chamber
On 15 April 2025, the Legislative Chamber held the first reading of the draft law. According to Gazeta.uz, deputies emphasized that:
- the bill was deliberately framework-oriented;
- the goal was not to limit technological innovation, but to introduce transparent rules and guard against abuses;
- deepfakes, misinformation, and human-less automated decision-making were considered the most urgent threats.
Following the first reading, the bill went through a period of refinement. UzDaily reports that by the time of the second and third readings, several clarifications were introduced:
- strengthened human rights protections;
- clearer wording of the "no decision solely by AI" rule;
- codified administrative liability for misuse of AI and illegal handling of personal data.
On 13 August 2025, the Legislative Chamber approved the bill in the third reading and forwarded it to the Senate.
4. The Senate Stage: Adoption and Political Justification
When the Senate considered the bill, it took the form of:
"Law on Amendments and Additions to Certain Legislative Acts of the Republic of Uzbekistan in Connection with Regulating Relations Arising from the Use of Artificial Intelligence Technologies."
This is important: Uzbekistan opted not to create a standalone "AI Act" but instead amended the existing Law "On Informatization" and related acts.
On 1 November 2025, the Senate approved the law during its 11th plenary session. Media reports (Yuz.uz, Fergana, Frank. uz) highlight the following positions articulated during the Senate review:
4.1. International benchmarking
Lawmakers stated that the drafting process considered the experience of:
- the United States, Canada, Russia, China, Japan, South Korea, and EU member states.
4.2. Economic competitiveness
Several senators argued that, with global expectations of USD 20 trillion in economic value from AI technologies by 2030, Uzbekistan needed timely regulation to avoid falling behind.
4.3. Alignment with the AI Strategy and national pilot projects
Senators referred to:
- the adoption of the AI Strategy;
- more than 30 pilot AI projects underway in public and private sectors;
- Uzbekistan’s rise by 17 positions in the AI readiness index, reaching the highest score in Central Asia.
4.4. Ethical and educational concerns
Some members expressed concern over students using AI tools to write assignments, potentially undermining independent thinking. Deputy Minister of Digital Technologies Rustam Karimjonov stressed that AI should be treated as a tool, not a substitute for human reasoning, and that ethical and digital literacy education was essential from early school years.
4.5. Senate’s final assessment
The Senate’s official communication emphasized that the law:
- promotes the development of AI technologies;
- establishes general principles for their use;
- reinforces the protection of citizens' rights and freedoms in AI-related relations.
5. Key Provisions of the Adopted Law
Although the law is a set of amendments, its core provisions form a coherent regulatory framework.
5.1. Definition and guiding principles
The law introduces an official legal definition of AI as technologies capable of analyzing data, learning, and assisting decision-making without direct human input, and it outlines the principles of state policy in this field.
5.2. Human oversight and prohibition of harm
The law prohibits:
- the use of AI systems or AI-generated information resources that could harm life, health, dignity, or other fundamental rights;
- making decisions affecting individual rights solely on the basis of AI-generated outcomes.
This codifies a form of human-in-the-loop requirement similar to emerging global standards.
5.3. Mandatory labelling of AI-generated content
All content produced with the help of AI—images, videos, audio, text—must be clearly labeled. The principal aim is to limit deception and reduce the circulation of deepfakes.
5.4. Personal data and administrative liability
The law introduces liability for:
- unlawful processing of personal data using AI;
- misconduct involving dissemination of such data online or through mass media;
- failure to comply with labelling requirements.
According to regional media (e.g., Liter. kz), penalties may include:
- fines of 50−100 basic calculation units;
- up to 15 days of administrative arrest;
- confiscation of equipment used in violations.
5.5. Institutional structure
The amendments identify a specially authorized national body for AI development, tasked with:
- attracting investment into AI;
- creating technical infrastructure, including data centers and AI-based government platforms;
- overseeing specialist training, retraining, and professional development.
Together, these measures connect regulatory oversight with the country’s broader digitalization agenda.
6. Expert Evaluation: What This Model Means
Uzbekistan’s regulatory approach can be described as:
- framework-based, not risk-tiered;
- integrated into existing legislation rather than standalone;
- centered on rights, transparency, and accountability.
Strengths
- Quick implementation through amendment of the Informatization Law.
- Strong emphasis on human rights safeguards and human oversight.
- Clear rules regarding synthetic content in response to rising deepfake risks.
- Direct alignment with the national AI Strategy and infrastructure development.
Challenges and open issues
- A framework law leaves considerable discretion for future by-laws, which may vary widely in strictness.
- Enforcement of content-labelling rules may be particularly difficult, especially across decentralized online platforms.
- The provisions on personal data require a coordinated institutional capability to prevent misuse.
- The legislation does not yet include a detailed, EU-style risk-based classification of AI systems, so future regulatory expansion is likely.
