Acquisition, adaptation, and improvement. What do ICT Week Uzbekistan experts recommend to keep pace with the development of AI?

Adopt, adapt, and advance. What do ICT Week Uzbekistan experts recommend to avoid falling behind in AI development?

At the Digital Education forum held as part of ICT Week Uzbekistan, representatives of international organizations and technology companies shared their views on the impact of artificial intelligence on education and the labor market.

At a time when the AI sector is expanding rapidly, the risks and opportunities it poses for countries and economies were at the center of the discussion.

According to Cecilia Paradi-Guilford, Senior Digital Development Specialist at the World Bank, to avoid falling behind in AI progress, countries must start applying the technology now, continuously adapt it to local needs and conditions, and selectively invest in advanced scientific research.

Meanwhile, Larisa Witcroft, representing Coursera, emphasized the importance of prioritizing people over infrastructure when adopting and investing in AI.

Inna Ozjogina, Regional Manager at AWS (Amazon Web Services), listed four conditions necessary for the proper use of artificial intelligence and for it to create more opportunities than threats. According to her, AI must be safe, reliable, accessible, and people must have sufficient knowledge to use it.

Cecilia Paradi-Guilford, Senior Digital Development Specialist at the World Bank, noted that today the main question is not whether artificial intelligence will destroy or create jobs. The key issue is which tasks will change, who will benefit, and who is at risk of being left behind.

In her view, any job consists of a series of tasks. Therefore, the fact that a specific job is affected by AI does not mean it will disappear completely; rather, AI may take over certain tasks inherent to it.

"A job is a collection of tasks. AI can take over some of them, help people work more productively on others, and create new tasks. The mere fact that a job is affected by AI does not mean it will disappear," Paradi-Guilford says.

According to a report, in economies similar to Uzbekistan, about 4.5 percent of jobs are at risk of automation. At the same time, the number of jobs whose productivity could increase due to AI is nearly four times higher.

"It follows that the risk for Uzbekistan is not the loss of jobs, but a missed opportunity: others—competitors—will reach this milestone faster, and the tools will develop before people have time to learn how to use them," the World Bank representative explains.

The report proposes a three-stage path for countries: adopt, adapt, and advance.

That is, starting to apply the technology now, continuously adapting it to local needs and conditions, and selectively investing in advanced scientific research.

"From a policy perspective, the most important issue in deciding which path to take is how to maximize the benefits to the economy and society," she says.

The World Bank representative listed four strengths of Uzbekistan: a high level of internet penetration, achievements in STEM graduates, significant investment in skills, and a recent sharp increase in the number of scientific publications on AI. However, the speed of adoption alone does not mean a result. Completing a task faster does not mean productivity is increasing in a company or that the economy is transforming. For this, AI must be applied widely and effectively.

Paradi-Guilford sees the most serious challenge for young people just entering the workforce.

"Based on US data, we did not observe large-scale job destruction or mass layoffs for young people entering the labor market, but there were certain delays due to lower hiring rates for young workers entering occupations more exposed to AI. This raises the question: how do people gain experience when entry-level, more routine jobs are changing, and these first jobs are truly important?" she said.

In other words, AI is now taking over the simple tasks that used to be performed by young, inexperienced specialists. It was precisely these tasks that served as a school of experience. The World Bank sees the solution to this in restructuring the internship and apprenticeship system.

"It is very important for young people to continue learning their field, to be able to use AI effectively, and also to understand when it makes mistakes. Consequently, this requires core human skills: critical thinking, communication, and the ability to learn continuously. Therefore, I think this clearly highlights the need to rethink key entry-level jobs, internships, and apprenticeship programs that will continue to develop these skills and lay the foundation for our future talent," Paradi-Guilford said.

From this perspective, the World Bank representative noted that AI literacy is necessary for everyone. Apprenticeship and internship programs should genuinely encourage learning through practice. To achieve this, the private sector should not just hire talent, but collaborate in creating this talent pool.

"I am not worried about AI making people redundant; I am worried that we will not adapt in time," she concludes.

According to Larisa Witcroft, AI is both a threat and an opportunity at the same time. Importantly, it is people, not the technology itself, who determine whether it becomes a threat or an opportunity.

Speaking about the threat of AI, particularly to the labor market, Witcroft cited some figures. She stated that since 2022, demand for junior specialists has decreased by 28 percent. The reason is that simple and repetitive tasks previously assigned to young employees are now performed quickly and cheaply by AI. Furthermore, today professional skills become obsolete in just 18–24 months. This means that the knowledge and skills acquired by young people quickly lose value, and at the initial stage, which should be the first school of experience in their professional career, they face the risk of competing with AI.

At the same time, Witcroft noted that AI is creating immense value.

According to estimates, the value created by AI will reach $22.5 trillion by 2031. At this point, she draws attention to the distribution of this value.

"Look at other figures: 74 percent of this value goes to just 20 percent of organizations. This means that the main issue and threat here is not the arrival of AI, but its uneven adoption. Entire countries risk being left behind. So, the question is what needs to be done to join that 20 percent group," she says.

Witcroft noted that the answer lies not in technology, but in people. Research conducted by the Boston Consulting Group (BCG) showed that 70 percent of the benefit from AI comes from people and processes. Only 20 percent is derived from technology, and 10 percent from algorithms. However, looking at companies' readiness for AI, the readiness level of technology and infrastructure stands at 43 percent, while for employees this figure is only 20 percent.

"This means we are investing twice as much in infrastructure and somewhat neglecting people. In reality, for AI to be effective in practice, people must use this infrastructure," she says.

At this point, she draws attention to two aspects—the extent to which the education system and business are ready for the wide implementation of AI in the country. When it comes to the education system, Witcroft believes that focusing solely on placing graduates in jobs is narrow thinking. In her opinion, graduates should not only be employable but also capable of creating something themselves.

Based on this, she concludes by sharing three recommendations. First, schools and universities should teach pupils and students not just the tool, but the entire cycle. By "cycle," it is meant teaching the student not simply to use an AI tool, but to test it, evaluate it, correct errors, and achieve a result.

Second, employers must not forget that 70 percent of the benefit from AI depends on people and processes, and should invest in training employees accordingly.

Third, it is advisable for governments and founders to invest in the application layer of AI, i.e., products designed to solve specific problems based on ready-made models. This is because this market is not yet fully occupied.