Russian without borders: learners from 114 countries study the language via an AI platform from RUDN University

Focus on pronunciation

Developed by RUDN University, the digital platform is designed to teach Russian to international learners. Over five years, it has attracted more than 15,000 users from 114 countries. Its defining feature is the use of artificial intelligence to tailor learning to each individual. Algorithms analyze a learner’s proficiency level, pace of progress and recurring mistakes, then generate a personalized learning path. This approach aligns with the expectations of modern learners, who increasingly prefer flexible, adaptive formats over one-size-fits-all courses.

A strong emphasis is placed on spoken practice. Built-in speech recognition enables users to train pronunciation and receive immediate feedback without a teacher’s involvement.

A major advantage is the mobile version with offline access. For learners in countries with slow or unreliable connectivity, this is not merely a convenience but a prerequisite for participation. In this way, the digital format expands access to Russian language education far beyond traditional academic centers.

Learning through mistakes

“We collected a thousand examples where errors are not only corrected but also explained through the relevant rules of the Russian language,” says Alexey Sorokin, senior researcher at the MSU Institute of Artificial Intelligence. Alongside the development of RUDN’s language platform, the MSU Institute of Artificial Intelligence is contributing parallel research in this field. Russian researchers there have proposed a practical way to train neural networks to handle some of the most challenging rules of Russian grammar.

The system studies individual students, understands where their knowledge gaps are, and recommends additional materials or training tools. The second AI component is speech recognition, because correct pronunciation is a core skill in learning Russian. Now students can improve their pronunciation skills at any time without a teacher’s involvement

Naim Rezaian

Head of the Artificial Intelligence Laboratory at RUDN University

At the core of the approach is a carefully curated dataset of typical mistakes involving complex spelling, punctuation and grammatical norms. In total, the sample covers 48 difficult rules of Russian – precisely those where both humans and algorithms most often struggle. Using these examples, AI systems are trained not just to spot inaccuracies but to understand why they occur and which rule has been violated.

Another outcome of the collaboration between linguists and engineers is a method for selecting texts with similar types of errors. This allows the neural network to see not isolated cases but entire classes of recurring mistakes, resulting in more accurate and meaningful corrections. As a result, AI learns to work with language logically rather than mechanically. Testing showed that the new methods increased the accuracy of correcting complex errors in Russian-language texts by 5–10 percent.

Linguistics meets code

In recent years, AI has become a standard tool for language-learning platforms. Major international services such as Duolingo rely on algorithms to personalize lessons and analyze learner behavior. EdTech platforms more broadly are actively deploying machine learning for adaptive content delivery, a trend also observed in educational ecosystems like Coursera.

Research indicates that personalized feedback and adaptive exercises improve motivation and learning outcomes compared with traditional methods. The RUDN project fits squarely within this global trajectory, applying similar mechanisms to Russian language education.

Beyond connectivity

The global digital language-learning market is growing rapidly. Valued at $12.49 billion in 2024, it is projected by analysts to reach approximately $35.68 billion by 2034. These figures reflect a broader shift toward online formats and personalized learning. Against this backdrop, projects that combine AI, mobility and offline access gain an advantage in the international education landscape, where demand increasingly extends beyond regions with high-speed internet.

The geographic reach of the user base points to sustained interest in digital formats for learning Russian. The platform can be integrated into university distance-learning programs and used in initiatives aimed at promoting the language abroad. For the IT sector, it serves as an example of applied AI in education, where value is created not through technological complexity alone but through usability and accessibility.