Як створити сильну ШІ-екосистему

22.10.2026, 09:20◉ 03 min readEditorial team “SHCHO STALOSIA”

В Україні дедалі більше спеціалістів із штучного інтелекту, але лише 2% з них працюють у публічному секторі. Щоб амбіції держави щодо ШІ стали реальністю, потрібна екосистема, яка впроваджує розробки.

English translation is temporarily unavailable. The Ukrainian original is shown.

However, the implementation of this ambition depends not only on the rules and state programs. According to AI House research, there are about 6,100 AI/ML specialists in Ukraine – 17% more than two years ago. At the same time, only 13% of them work with AI outside of IT, and 2% in the public sector.

This gap shows that the problem is not so much the number of specialists, but how to turn their expertise into practical solutions. This requires an ecosystem that connects the AI community with the tasks of business and the state, gives access to data and research, and helps bring developments to implementation.

This is startup thinking: first problem and hypothesis testing with all stakeholders, then resources, team and product. If one of these links falls out, the implementation of the technology will take resources, but will not produce results.

Details

For example, the Canadian ChartWatch system was created for the specific need of the hospital: to identify patients at high risk of a sharp deterioration in their condition in advance.

The ecosystem needs a mechanism by which each new task follows the same clear path: from assessing the problem and data to team building, development, testing and implementation of the solution.

Context

By 2030, Ukraine aims to become one of the three world leaders in the level of AI implementation in the public sector. In 2026, the Ministry of Digital Development began to prepare a profile law. This was preceded by the launch of a regulatory test environment (sandbox), where companies test their products under the supervision of experts.

Every hour, the system analyzes more than 100 indicators and warns doctors. It was developed with doctors, patients and hospital management. After its implementation at Toronto St.

What's next

One of the main mistakes in working with AI is to start with technology. Before engaging engineers, it is worthwhile to find out who will use the solution, what result is expected from it, what data is available and how success will be measured. The task must have an owner interested in the further use of the result.

The winning decision is already being introduced into judicial practice and is planned to be introduced into state digital and archival processes. Other effective models are used to mark up descriptions of the state archive funds. Agreed cooperation with the State Judicial Administration on the recognition of court cases and applications.