“Artificial intelligence is becoming a strategic infrastructure” — ICT Week Uzbekistan experts

“Artificial intelligence is becoming a strategic infrastructure” — ICT Week Uzbekistan experts

State policy in the field of digitalization, practical examples of introducing artificial intelligence into business, and the risks of monopolization of the technology market were discussed at the Future Intelligence panel session held within the framework of ICT Week Uzbekistan.

The discussion was attended by representatives of relevant government agencies, international technology companies, local startups, and the banking sector.

The First Deputy Minister of Digital Technologies announced that the Ministry of Digital Technologies, in cooperation with Huawei and the Global Computing Consortium, is preparing the Uzbekistan AI White Paper document.

Four years ago, a similar document — the “E-Government White Paper” — served as the basis for the roadmap for the development of public services.

The new “White Paper” aims to develop AI in the country, transitioning from individual initiatives to systematic implementation, in accordance with the approved strategy for the development of AI technologies for the period up to 2030.

The document provides an objective assessment of the national ecosystem, identifying its strengths and limitations, and also points out the areas where technologies can yield the greatest practical effect.

In addition, the AI White Paper can become an opportunity map for technology companies and foreign investors, as it allows them to see a “clear picture” — state priorities, the state of infrastructure, promising sectors, human resource potential, and long-term directions of market development.

“At the same time, it is important for us to obtain not just a set of general recommendations, but a working document that will help government agencies in making decisions, business in finding new directions of development, and investors in seeing specific opportunities. That is precisely why the project includes sectoral cases, infrastructure analysis, and results evaluation tools,” Pekos concluded.

The development of the Single Interactive State Services Portal (MyGov) and how AI will change citizens' interaction with the state were also discussed.

According to him, over the past year, the platform has received more than 66 million applications, and more than 250,000 people use it daily. The MyGov mobile application has been installed by more than 16 million users. There are 667 services available in the application, which constitutes about 72 percent of electronic public services; there is no longer a need to visit offices to use them.

Abdufattohov noted that the ecosystem has completed the stage of consolidating departmental services into a “single window” and is moving towards services that take into account human life situations.

“Today, based on the ‘three steps’ principle, more than 200 services have been simplified. More than 130 composite services combining several procedures of different departments have been introduced. 21 proactive services have been launched — where the state offers the service to the citizen itself, without waiting for their application. More than 150 certificates, which previously had to be collected by visiting office after office, have been abolished. All this works thanks to integration with more than 160 information systems. This is the technological foundation of interdepartmental cooperation,” he said.

The platform's plans for the near future include launching three AI-based projects: the MyGov AI Assistant voice and text assistant, the MyGov Speech to Text voice application filling system, and the Document AI document recognition service.

It is also planned to create another 80 proactive and composite services, shorten the delivery time of 80 services, and reduce the cost of 25 public services. At the same time, 20 of the most in-demand procedures will be transferred to an online format through the portal — requiring no personal visits.

In parallel, traditional public service centers will also change. They will be converted into a digital space format, where citizens can use services independently through terminals, while employees will be assigned a mentoring role.

Since the establishment of the Alliance for the Development of Artificial Intelligence in 2025, 40 organizations, including government agencies, private companies, and universities, have joined it, the Director of the AI Alliance announced.

He explained that the organization's task is to unite market participants in a single space and create opportunities for them to share experience.

“No single sector can develop this field alone: government agencies define the regulatory framework, private business acts as the driver of implementation, and universities provide the sector with personnel,” Muhitdinov said.

According to him, the organization is focusing on practical initiatives.

Under the “Five Million AI Leaders” program, about 1,000 young people from the regions improved their skills in this field this year.

Another important project was the National AI Hackathon, held in 11 regions of the country. Muhitdinov noted that in order to showcase the potential of the regions, the Tashkent stage was deliberately postponed to the final. The competition brought together more than 3,000 participants who prepared about 200 prototypes, and after the final, some of the teams received investments in the amount of 20,000–30,000 dollars from venture funds.

Yefei Sun announced that the Alipay ecosystem, built on the super-app principle, has expanded to 4 million mini-programs today.

To facilitate navigation in the suite of services, the company is changing its approach to user experience — the market is moving from traditional button and menu interfaces to AI agent models. With this approach, the customer no longer needs to search for the necessary sections independently — interaction with the platform takes place through natural communication with a digital assistant, which performs the necessary actions itself.

Anvar Narzullayev spoke about the introduction of a fully automated system for job preparation. All stages, from technical interviews and case solving to video interviews analyzing the candidate's behavior, are conducted using language models. In general, more than 70,000 graduates have passed through the platform to date.

The company achieved the most significant effect from the introduction of AI in internal processes. According to Narzullayev, the Mohirdev developer team was reduced to two people: a team lead who approves the technical plan and a quality assurance (QA) engineer who accepts the result. AI generates all the code for frontend and backend.

As a result, the time for developing new features was reduced by 67 times — the average period decreased from 14 days to 4–6 hours. In addition, a single data repository was created in the company with the ability to execute analytical queries via MCP, integrating CRM, financial systems, and legal documents.

Spartak Tetrashvili called for lowering the level of emotional expectations from the term “artificial intelligence.”

“Artificial intelligence, in my opinion, is a marketing name, because it does not exist [in the classical sense]. What we call that today are extremely powerful models capable of processing data very efficiently,” he said.

Tetrashvili compared the emergence of AI to the introduction of computers 50 years ago, when the financial sector was transitioning from manual calculation to automation.

TBC Bank highlights three main scenarios of interaction with customers based on AI:

To automate internal processes, a separate “AI-ization” department was established in the bank, and the next main direction will be AI agents. Over the past three years, the bank's investment in these technologies amounted to at least 15 million dollars, Tetrashvili said.

“[This] is not a small sum, but we are definitely seeing great benefits. Perhaps this benefit is not directly reflected in the numbers, but this is the future. If you don't buy a computer today, you won't understand how to use it tomorrow. Therefore, we believe we must be where novelty and innovation are,” he concluded.

Ruslan Namozov emphasized that the company focuses only on the business effect coming from the introduction of technologies. For e-commerce, search and personalized recommendation algorithms, as well as generative models for content moderation, represent the greatest value.

According to him, all AI solutions in the company undergo mandatory marketing testing, as the main risks manifest themselves when scaling products.

“Everything can be filled with AI tokens, but if it is not effective, there is no point in it. The second problem is data security. Solutions working on a small scale and solutions working on a large-scale flow are completely different things, because all unexpected situations must be taken into account,” Namozov said.

He noted that AI is changing traditional principles of education. Previously, an entire class of 30 people listened to the same teacher, but now AI allows the introduction of personalized learning.

“With the help of AI, we can finally provide personalized knowledge to each person. These 30 people can sit in one classroom, but receive information in a form that suits them exactly. A student interested in football can learn the same law of physics through examples taken from football, and a student interested in music can learn it through music,” he explained.

Speaking about training IT specialists, Abdullayev highlighted two main aspects. First, business updates the AI stack every few weeks, so universities must quickly adapt their curricula. To this end, an Industry Advisory Board consisting of heads of IT companies and experts has been established at Inha University.

Second, employers are increasingly refusing to hire junior developers and are lowering the demand for “middle” level specialists, which is why universities must train personnel who can not only write code but also manage the process of creating finished products using AI.

As an example of adapting training, the speaker cited a practical case at the university: students are given an assignment that usually takes a month to develop, but are asked to complete it in a two-hour session. The main conditions are active collaboration in a team and mandatory use of AI tools.

Farrukh Qoraboyev devoted his speech to the issues of market regulation in the era of AI, as well as technology monopolization and antitrust policy.

“The main issue today is not who controls AI, but who controls the markets where decisions are already being made by AI,” he said.

As confirmation of this thesis, Qoraboyev cited a study by Harvard University and Pennsylvania State University published in 2024. During the experiment, two AI agents that did not have a direct communication channel with each other independently set prices in order to maximize profit. As a result, they agreed on highly competitive prices, and when aiming for long-term profitability, the total profit approached the monopolist level. At the same time, classical mechanisms of antitrust control do not work in this case, because there was no fact of legal collusion or correspondence between the algorithms.

Qoraboyev also noted an unprecedented concentration of resources among leading tech giants. In 2026, Amazon, Alphabet, Microsoft, and Meta plan to spend a combined total of approximately 725 billion dollars on capital expenditures, a large part of which is related to the development of AI infrastructure. This is 77 percent more than last year's figure.

Answering the question about the role of the state, Qoraboyev called for refraining from hastily adopting a separate law on AI. He cited the European Union as an example, where the strict AI Act restricted the local technology sector, and as a result, European companies began to lag behind competitors in America and China.

Instead of strict regulation, the representative of the Competition Promotion Committee proposed strengthening related legislation.

Regarding antitrust control, Qoraboyev favored supplementing the competition law with the concept of “quasi-mergers” — these are “license plus team” deals, with the help of which technology corporations bypass standard clearance procedures.

In the field of pricing algorithms, he proposed obliging companies to train systems only on data older than 12 months, and giving the antitrust authority the right to request algorithm logs in concentrated markets such as fuel, air tickets, and cement.

According to Qoraboyev, the transfer of citizens' personal data from public services for training models should be carried out only with the explicit consent of users.

The official also considers the development of sovereign AI important. For the public sector, instead of large and expensive systems, he proposes creating small modular data centers with compact models obtained by distillation.

In the context of the establishment of the World Artificial Intelligence Cooperation Organization (WAICO) in Shanghai, he drew attention to the geopolitical role of technology.

“AI is no longer just a technological innovation. It is becoming a strategic infrastructure like electricity, the internet, and space technologies, as well as a key area of geopolitical competition between leading powers, primarily the US and China,” he said.

Uzbekistan, along with Kazakhstan, Kyrgyzstan, and Tajikistan, was among the 29 founding countries of WAICO. According to Norov, the co-founder status gives the countries of the region the opportunity to participate in the development of global rules, decision-making mechanisms, and technical standards, rather than adapting to the norms of others.

The speaker emphasized that the new organization is based on mutual trust, respect for sovereignty, and consensus-based decision-making. According to him, for Uzbekistan, which strives to become a regional hub, membership in WAICO opens the way to developing data processing centers and cloud platforms, attracting investments, and establishing joint research centers.