SELF-LEARNING ENGLISH WITH AI – PRACTICAL AND ECONOMIC ASPECTS

Abstract

This paper explores the potential of using artificial intelligence (AI) for self-learning
and self-practising English among university students of economics, law, and engineering.
However, its findings can be readily applied to students of other non-linguistic disciplines, as well
as to learners whose average command of English is below CEFR level C. The paper offers an
overview of popular AI systems and is valuable for several reasons – it presents the AI
applications available to both students and adults; highlights their advantages and limitations,
and thus assists all potential and current users in deciding which to choose. The research is
based on two data sources. The first is a Google form survey completed by students of the Kosovo
and Metohija Academy of Vocational Studies in Leposavić (Peć Department), the Academy of
Professional Studies Šumadija in Kragujevac (Trstenik Department) and the Faculty of Law of
the University of Priština, temporarily seated in Kosovska Mitrovica during the summer semester
2025. The second source is the evaluation of the AI systems conducted by the authors of this
paper. They assessed their qualities and identified the language skills and the level of language
knowledge each application focuses on. The results indicate that students are well-acquainted
with AI applications, with 70% reporting frequent or occasional use. AI tools are used for
learning, entertainment and work. Duolingo, Busuu and Promova ranked highly for vocabulary
learning and speech practice, and Grammarly and Grammar Up for giving spelling, grammar
and writing style guidance – key components of effective English communication. AI applications
also function as conversation partners and support non-traditional language practice, thus
helping to reduce the anxiety of speaking English in public. In addition, the study identifies
certain limitations that can hinder the user's progress – technical difficulties, lack of routine in
use and concerns about the future development of AI, which may implicitly lead to the
dehumanisation of the learning process. In terms of economic viability, AI-assisted English
learning is cost-effective – learners using AI invest time rather than money. They choose the
high-demand skills in the market, and thereby acquire knowledge that increases their own
market value and competitiveness. Access to knowledge becomes broader, and the overall
learning costs are lower. This is especially relevant for individuals with limited educational
opportunities and for marginalised groups, contributing to greater social and economic
inclusion. However, AI-based learning reveals some hidden challenges and uncertainties
regarding the sustainability and quality of the knowledge acquired. It is therefore unlikely that
AI will ever fully replace human teachers. It cannot replicate classroom traits such as emotional
intelligence, adaptability to different needs, creativity, and the ability to develop interpersonal
relationships. Instead, AI can be viewed as a powerful supplement and support for teachers,
enhancing effective and engaging learning for everyone. The optimal strategy for ensuring
quality and sustainability is to combine AI tools with expert guidance and a structured learning
framework. Therefore, teacher education should adapt to these technological trends and include
AI training, while also addressing its limitations, risks and challenges. Consequently, AI should be
embraced as an ally in learning, not a rival. A key prerequisite is that AI is used ethically,
respecting privacy and equal opportunities for all students.
Key words: English language, AI, advantages, disadvantages, applications