The phrase “Lend me an agent until Friday” may come to sound as familiar in student dormitories as the words “Hey, got the notes?” once did in the popular Soviet film comedy Operation Y.
At the end of October, Professor Sergey Zverev of HSE University told RIA Novosti that by 2028 universities will graduate not only a specialist but also their AI double. Zverev heads the School of Communications at the HSE Faculty of Creative Industries. The remarks were made at the Moscow Department of Information Technology's business space “MoskHab.Skolkovo.”
Learning in parallel with a double
According to the expert, the familiar “teacher — student” scheme is changing: when both students and teachers use neural networks, checking standard homework loses its point. He considers fighting algorithms pointless.
Instead, Zverev proposes multi-agent learning. On enrollment, a student creates a digital double, and it learns alongside them. The person solves tasks, checks the AI's results, corrects mistakes and passes knowledge on to it. The double must think like a specific student, not like a universal chatbot. After graduation, the person receives a diploma and a set of AI tools familiar with their approach to work.
Zverev names critical thinking as the professional's key skill. The conclusion is formulated thus: “A person with artificial intelligence will replace a person without it.”
Neural networks are already in the classroom
According to a study by the Financial University cited by TASS, in 2024–2025 the share of students using neural networks grew from 66% to 92%.
The OECD report Digital Education Outlook 2026 notes that general-purpose AI tools improve learning outcomes in the short term. AI tutors built on the principles of active learning can also produce good results.
What already exists in the world
The Technical University of Munich is testing a teacher's double in the “Professor Digital Twin” project. The American Bloomwards Academy offers the “AI Co-Pilot” model, in which students themselves train a personal AI. These initiatives are still at the stage of development and experimentation.
A diploma and a digital sidekick
Previously, a graduate left university with a diploma and a stack of notes. By 2028, if Zverev's statement comes to pass, they may leave with a diploma and their own digital sidekick, which has learned to think the same way they do. More than 90% students already work with neural networks, and universities around the world are trying various models: from a teacher's double to a personal AI assistant for every first-year student. The question is not whether students will get an agent, but what it will learn from its master. The diploma will still go to a human, but the Friday when the agent has to be returned may become the most interesting day of the semester.
