Universities have strengthened the control of theses against the background of the development of AI. What is known

AI detectors do not always allow us to unambiguously establish the authorship of student papers, and their use in educational practice is controversial. Experts and government officials point to the need for a comprehensive assessment and adaptation of verification approaches. All the details are in the Izvestia article.

Goals and objectives: agencies have established standards for the implementation of neural networks

Why has the bar been set low for business and how will efficiency be determined?

What is known about checking diplomas for AI

Alexander Tolmachev, deputy chairman of the State Duma Committee on Youth Policy (United Russia faction), told Izvestia that the spread of artificial intelligence creates a new situation in the field of intellectual work. According to him, correct and logical texts are increasingly being attributed to the results of neural networks without sufficient justification. Nevertheless, in such cases, the student has the right to challenge the conclusions of the teachers by providing a list of sources and previous work.

In addition, he noted that universities can take into account the student's academic history when assessing the authorship of a diploma. If the student systematically showed poor results, the teachers may have doubts about the independence of the extensive work. At the same time, linguistic expertise remains one of the verification tools, but it requires comparison with other texts of the author.

According to the deputy, the current legislation does not yet contain specialized standards for evaluating AI work, but it is gradually being supplemented. It is reported that the State Duma and the government are working on regulating the digital sphere, including issues of cybersecurity and the platform economy.

The attraction of tycoons: the high-profile trial on AI ethics turned out to be a redistribution of the chatbot market

What lies behind Elon Musk's altruistic statements about protecting the interests of humanity

What rights do students have?

Lawyer Ilya Rusyaev believes that the automatic report of the anti-plagiarism service cannot be the only reason for not being allowed to defend. According to him, federal legislation does not provide for a direct dependence of access on the percentage of machine generation. The decision should be based on an objective and independent assessment.

In addition, he noted that the student has the right to challenge the university's decision through internal commissions or government agencies. Among them are the dean's office, the rector, as well as Rosobrnadzor and the prosecutor's office. Experts note that the court remains the final instance, where it is possible to demand recognition of the decision as illegal.

It clarifies that the university is required to provide not only the percentage of verification, but also a full report with text analysis and expert opinion. Local regulations governing the use of AI are also taken into account.

Why are the detectors malfunctioning and what is changing?

Daria Rutsinskaya, IT specialist and Director of Marketing at Zapusk Group, noted that AI detectors work on the principle of searching for typical language patterns, which leads to false positives. The academic style often coincides with the features of machine generation, since both systems rely on similar sources. According to international research, the accuracy of such instruments remains limited and unstable.

At the same time, it is reported that the massive use of neural networks by students has already become a common practice. At the same time, some universities are moving from prohibitions to a model of declaring the use of AI. This approach assumes the responsibility of the student for the content of the work, while openly indicating the tools used.