Larger Groups of Students Use AI More Effectively in Learning
Researchers at the Institute of Education and the Faculty of Economic Sciences at HSE University have studied what factors determine the success of student group projects when they are completed with the help of artificial intelligence (AI). Their findings suggest that, in addition to the knowledge level of the team members, the size of the group also plays a significant role—the larger it is, the more efficient the process becomes. The study was published in Innovations in Education and Teaching International.
Group projects are an essential and common part of higher education, but there is still uncertainty about what makes teamwork effective.
The situation has become more interesting since the emergence of AI, which students have started to actively use in their studies. Experts from HSE University, including Galina Shulgina, Aleksandra Getman, Ilya Gulenkov, and Jamie Costley, explored how the characteristics of groups—the size and level of participants’ knowledge—affect the outcomes of work when AI is involved.
The study included 196 second-year undergraduate students, 55% of whom were male and 45% female. They had to solve problems as part of a team in a 16-week macroeconomics course. The students were divided into groups of five to eight people with varying levels of knowledge and experience. At first, the students worked independently on tasks. Then, they attended four seminars where they used ChatGPT 3.5 as a group tool. The goal was not simply to receive an answer from the AI, but to critically analyse it, apply economic models from the course, and present a comprehensive solution.
Researchers evaluated the quality of solutions based on the accuracy and detail of students' responses. Teams that not only used AI correctly but also revealed its limitations earned the highest scores, demonstrating a deeper understanding of the material.
The scientists identified several patterns in the use of AI by groups. Firstly, the best results were achieved by teams with members of similar levels of expertise. However, teams with a wider range of knowledge often performed less effectively. This is despite the fact that, in pedagogy, it is often believed that diversity of knowledge can help rather than hinder a team's performance.
Galina Shulgina
‘We were surprised to discover that the wider the range of student grades, the lower the quality of the final decision. This may be because the more prepared students spent time discussing and reaching an agreement on a solution, rather than focusing on the task itself, while less prepared students were unable to fully utilise the AI capabilities available to them. More skilled students are better at interacting with AI, as they can formulate more complex queries, critically evaluate the responses, and use this information to reason through problems,’ explains Galina Shulgina, junior researcher at the International Laboratory of Research and Design in eLearning at HSE University.
Secondly, the data showed a clear positive correlation between a larger team size and better performance when working with AI. Larger teams, with seven to eight members, performed better on average compared to teams with five to six members. Each additional member contributed to the final score, contrary to the common belief in pedagogy that smaller teams are more effective. Scientists argue that larger teams have more intellectual resources and a variety of perspectives, which help them interact more productively with neural networks.
Aleksandra Getman
‘However, this does not mean that efficiency gains will continue infinitely. After a certain point, negative effects may start to appear, such as difficulty in coordination and increased time to coordinate and maintain shared understanding of the task,’ explains Aleksandra Getman, junior researcher at the International Laboratory of Research and Design in eLearning at HSE University.
Despite the need for further research, the authors believe that in order to optimise the use of AI in education, students with similar educational levels should be grouped together in large classes. The researchers suggest that AI could be applied to the study of any subject.
Ilya Gulenkov
‘There is a potential for incorporating AI into group work in any course, regardless of the field of study or level of training. The key task of the teacher in organising such work is to set students’ expectations in advance about how and why AI can be used in their coursework. If students see examples of successful application of AI, then it can become an additional team member in any subject. We observe how students are using more advanced versions of the models (ChatGPT 5, ChatGPT 5 Thinking, etc), and we see great potential for student–AI collaboration. This applies not only to simple, standardised tasks, but also to complex ones that require in-depth understanding, working with multiple sources, and advanced reasoning. The role of students' own expertise in interacting with these models is becoming increasingly important. All models now provide plausible answers, but it is essential to critically evaluate their content,’ says Ilya Gulenkov, lecturer at HSE University’s Faculty of Economic Sciences.
See also:
IDLab: Fascinating Research, Tough Deadlines, and Academic Drive
The International Laboratory of Intangible-driven Economy (IDLab) was established at the HSE campus in Perm 11 years ago. Its expertise in data processing and analysis allows researchers to combine fundamental studies with applied projects, including the development of risk and cybersecurity models for Sber. The head of the laboratory, Professor Petr Parshakov, and Senior Research Fellow Professor Mariya Molodchik spoke to the HSE News Service about IDLab’s work.
HSE Tops Ranking of Universities Participating in Priority 2030 Programme
The Russian Ministry of Science and Higher Education has published an updated list of participants in the Priority 2030 programme. A total of 106 universities will receive support this year. HSE University was included in the first group and topped the ranking.
HSE Psycholinguists Launch Digital Tool to Spot Dyslexia in Children
Specialists from HSE University's Centre for Language and Brain have introduced LexiMetr, a new digital tool for diagnosing dyslexia in primary school students. This is the first standardised application in Russia that enables fast and reliable assessment of children’s reading skills to identify dyslexia or the risk of developing it. The application is available on the RuStore platform and runs on Android tablets.
HSE University to Join Physical AI Garage Project by Yandex
Yandex is collaborating with leading Russian universities to launch a new educational programme called Physical AI Garage. This initiative unites five universities—HSE University, ITMO, MIPT, MAI, and MEPhI—to train future professionals in physical artificial intelligence by tackling real-world industrial challenges. The programme is free, and participants will receive scholarships.
HSE Scientists Optimise Training of Generative Flow Networks
Researchers at the HSE Faculty of Computer Science have optimised the training method for generative flow neural networks to handle unstructured tasks, which could make the search for new drugs more efficient. The results of their work were presented at ICLR 2025, one of the world’s leading conferences on machine learning. The paper is available at Arxiv.org.
Physicists Propose New Mechanism to Enhance Superconductivity with 'Quantum Glue'
A team of researchers, including scientists from HSE MIEM, has demonstrated that defects in a material can enhance, rather than hinder, superconductivity. This occurs through interaction between defective and cleaner regions, which creates a 'quantum glue'—a uniform component that binds distinct superconducting regions into a single network. Calculations confirm that this mechanism could aid in developing superconductors that operate at higher temperatures. The study has been published in Communications Physics.
Neural Network Trained to Predict Crises in Russian Stock Market
Economists from HSE University have developed a neural network model that can predict the onset of a short-term stock market crisis with over 83% accuracy, one day in advance. The model performs well even on complex, imbalanced data and incorporates not only economic indicators but also investor sentiment. The paper by Tamara Teplova, Maksim Fayzulin, and Aleksei Kurkin from the Centre for Financial Research and Data Analytics at the HSE Faculty of Economic Sciences has been published in Socio-Economic Planning Sciences.
Mistakes That Explain Everything: Scientists Discuss the Future of Psycholinguistics
Today, global linguistics is undergoing a ‘multilingual revolution.’ The era of English-language dominance in the cognitive sciences is drawing to a close as researchers increasingly turn their attention to the diversity of world languages. Moreover, multilingualism is shifting from an exotic phenomenon to the norm—a change that is transforming our understanding of human cognitive abilities. The future of experimental linguistics was the focus of a recent discussion at HSE University.
New Models for Studying Diseases: From Petri Dishes to Organs-on-a-Chip
Biologists from HSE University, in collaboration with researchers from the Kulakov National Medical Research Centre for Obstetrics, Gynecology, and Perinatology, have used advanced microfluidic technologies to study preeclampsia—one of the most dangerous pregnancy complications, posing serious risks to the life and health of both mother and child. In a paper published in BioChip Journal, the researchers review modern cellular models—including advanced placenta-on-a-chip technologies—that offer deeper insights into the mechanisms of the disorder and support the development of effective treatments.
Using Two Cryptocurrencies Enhances Volatility Forecasting
Researchers from the HSE Faculty of Economic Sciences have found that Bitcoin price volatility can be effectively predicted using Ethereum, the second-most popular cryptocurrency. Incorporating Ethereum into a predictive model reduces the forecast error to 23%, outperforming neural networks and other complex algorithms. The article has been published in Applied Econometrics.