Artificial Intelligence in Uzbekistan: Legal Frameworks and Guidance for Investors (2026)

In recent years, Uzbekistan has made substantial progress in regulating digital technologies. For investors in the field of artificial intelligence (AI), this marks a transition from uncertainty to a framework of clearer—though increasingly demanding—rules. Below is an overview of the principal legal considerations to take into account when launching and scaling AI projects in the country.

1. Strategic Regulation and the “Regulatory Sandbox”

The regulatory basis of the current phase consists of state digital development programmes, including Presidential Decree No. UP-6079 (Digital Uzbekistan – 2030) and Presidential Resolution No. PP-255 aimed at advancing artificial intelligence technologies.

A significant step was the adoption of the Law of the Republic of Uzbekistan dated 21 January 2026 No. ZRU-1115 On Amendments and Additions to Certain Legislative Acts of the Republic of Uzbekistan in Connection with the Regulation of Relations Arising from the Use of Artificial Intelligence.

This law formally introduced:

the legal definition of artificial intelligence;
fundamental principles for its application;
the role of the state in infrastructure development and attracting investment.

One of the practical tools remains the regulatory sandbox, allowing businesses to:

test AI solutions in a controlled environment;
reduce regulatory risks at the launch stage;
engage with competent authorities (primarily the Ministry of Digital Technologies).

In practice, participation in such a regime requires prior legal preparation of the project.

2. Personal Data Localisation: A Key Requirement

The Law of the Republic of Uzbekistan On Personal Data (Law No. ZRU-547), including localisation provisions (Article 27-1), requires personal data of Uzbek citizens to be stored and initially processed within the territory of the country.

For AI projects, this means:

the need to host data on local infrastructure;
restrictions on using foreign cloud solutions without adapted architecture;
implementation of hybrid or fully localised IT models.

Since 2025, enforcement practice in this area has become significantly stricter.

Additionally, following the adoption of Law No. ZRU-1115:

direct administrative liability has been introduced for unlawful processing of personal data using AI;
oversight of online dissemination of such data has been strengthened.

Key risks include:

blocking of services;
fines;
restrictions when dealing with government bodies and major corporate clients.

Accordingly, investors are advised to conduct prior data architecture audits and consider local data centres.

3. Restrictions and Principles Governing AI Use

AI regulation in Uzbekistan is developing through a combination of “soft law” guidance and binding legal rules.

Core principles include:

prohibition of discriminatory outcomes;
informing users when AI is being applied;
promoting algorithmic transparency.

At the legislative level, mandatory restrictions now include:

AI systems must not violate human rights, freedoms, honour or dignity;
legally significant decisions must not be based exclusively on AI-generated conclusions.

This effectively creates a human-in-the-loop model, whereby final decisions must remain under human control.

4. Tax and Customs Incentives

A special regime for IT and AI projects continues through IT Park Uzbekistan.

Key benefits include:

exemption from most corporate taxes;
reduced personal income tax rates (including 7.5% in certain cases);
customs incentives for importing equipment, including servers and GPUs.

This regime remains one of the principal drivers of the jurisdiction’s investment attractiveness.

5. Intellectual Property and AI Products

Ownership of rights in outputs created using AI remains a critical issue.

Under the current legal approach:

copyright protection applies to works created by a human author;
AI is treated as a tool rather than an author.

Practical implications for investors:

contractual arrangements are of central importance;
rights to models, datasets and outputs should be separately regulated;
trained models and parameters (weights) should be structured as know-how where appropriate.

6. Practical Risks When Launching an AI Project

The most common challenges include:

non-compliant architecture from a localisation perspective;
absence of properly documented rights to models;
underestimation of personal data regulation;
use of AI in sensitive decision-making processes without human oversight.

These issues do not always prevent launch, but they can materially affect stability and scalability.

7. Practical Recommendations

Before launching an AI project in Uzbekistan, it is advisable to:

conduct legal and technical audits of data processing;
determine an appropriate localisation model (local or hybrid);
structure intellectual property rights;
assess eligibility for the regulatory sandbox;
consider obtaining resident status in IT Park Uzbekistan;
implement human oversight procedures for legally significant decisions.

Conclusion

The AI investment environment in Uzbekistan is developing as a model of controlled incentivisation: the state offers substantial benefits while simultaneously tightening business accountability requirements.

The key innovation of 2026 has been the formal legislative introduction of:

core principles for AI use;
limits on autonomous decision-making;
liability for violations involving data processing.

For investors, this means that successful market entry now requires comprehensive legal structuring from the earliest stages of the project.