Regulations defining the policy on artificial intelligence (AI) have been approved in only 14% of Russian universities and scientific organizations. In another 72% of organizations, such documents are absent. This is confirmed by a study by the Institute for Statistical Studies and Economics of Knowledge of the HSE University.

The survey involved representatives of 250 universities and 174 research institutes. In 65% of organizations, the use of AI in scientific work is already supported. The most common measures are seminars and master classes on the use of neural networks. Such events are held by 48% of universities and research institutes.

In 23% of organizations, AI advanced training programs have been introduced. Internal grants for research using AI are provided by only 10% of institutions.

What experts say

Alexander Bukhanovsky, head of the Institute of Artificial Intelligence at ITMO University, believes that many universities and research institutes have taken a wait-and-see approach. They need to accumulate experience in using AI to develop their own regulatory framework. The existing experience of teachers and students is very fragmented, as each of them uses different AI models, the expert explained.

Azamat Zhilokov, director of the MIPT AI Institute, stated that it will take time to agree on AI requirements. Since neural networks are used both for editing scientific text and for testing hypotheses, a different level of result control is needed. Universities and research institutes should know which tools to use to avoid errors and incorrect conclusions, he added.

Petr Ototsky, deputy director of the "Digital School of Public Administration" center at VSGU RANEPA, reminded that in 2025, AI regulations appeared in more than 10 universities in the country. The documents recommend teaching methods to teachers that prevent students from passing off the results of neural networks as their own. The approved concepts bring AI out of the "gray area" and decriminalize their use in the educational environment, the expert concluded.