The education system in Russia is facing an unprecedented challenge: artificial intelligence has ceased to be just a tool and has turned into a mass "author" of student papers. According to the "Antiplagiat" platform, in the second quarter of 2026, signs of neural network usage were detected in every third checked paper. This indicates that discussions about the harm or benefit of AI have been replaced by the need to develop specific control mechanisms.
Statistics: 12% growth in a year
Figures published by RBC demonstrate an alarming trend. During the period from April to June 2026, the service checked 2.73 million papers. Traces of AI generation were found in 983,700 of them (36% of the total volume). For comparison, a year earlier, this indicator was 12% lower. Moreover, if in 2025 the volume of probable generation within documents was 5.4%, by 2026 it had risen to 8%.
Experts note that this is not just statistical noise, but a systemic shift in how students approach completing academic assignments.
Which subjects are under attack?
The use of neural networks is unevenly distributed. Leaders in the number of papers "stuck" in AI are humanities and social sciences. Elements of generation were most often identified in texts on art history, media and mass communications, information science, medicine, and political science. Conversely, in technical and exact disciplines — veterinary sciences, mechanics, mechanical engineering, and mathematics — traces of neural network work are found significantly less often.
From prohibition to regulation
Evgeny Lukyanchev, Executive Director of "Antiplagiat," states that the attempt to completely stop the spread of AI in education is doomed to failure. In his opinion, the time for discussions is over, and the education system now needs concrete solutions. The main task is to stop simply recording the fact of AI presence in the text and start analyzing the nature, depth, and purpose of its use.
The company is already developing a new verification system that will allow universities to see exactly how a student used neural networks. Lukyanchev emphasizes that with the introduction of such technologies, higher education institutions will need clear rules: where the use of AI is permissible and where it is categorically prohibited. Such a comprehensive approach should become a new industry standard.
The end of the era of classic theses?
The transformation of the knowledge assessment system is being discussed at the ministry level. Earlier, the Russian Minister of Science and Higher Education, Valery Falkov, stated that the development of AI has led to the fact that classic written papers no longer reflect the real level of student training. In this regard, he suggested that writing theses might lose their relevance.
In the future, according to Falkov, only an oral answer, and not a text generated by an algorithm, will be able to objectively assess a student's real knowledge and competencies. This poses a task for teachers and university administrations to radically revise approaches to examination and project defense.