Introduction
Understanding the role of Artificial Intelligence (AI) in language acquisition is fundamental for multilinguals studying in Northern Cyprus. The current study examines Syrian students’ perceptions of learning Turkish, the effects of AI and contextual exposure, the challenges encountered in the acquisition process (grammar, vocabulary, speaking, and pronunciation), AI support for language skills, and their recommendations for using AI to learn Turkish efficiently. Thus, this investigation drew on Krashen’s input hypothesis, Swain’s output hypothesis, and the SAMR model.
Methods
This study used semi-structured interviews as a qualitative approach to collect data. The developed questions were organized into four main themes: perceptions, AI and contextual factors, challenges and AI support, and future suggestions. Due to exam schedules, female participants could not be reached; therefore, thirteen male pharmacy students, ranging from freshmen to seniors, were interviewed online in Arabic. The transcriptions were generated using digital tools (TurboScribe, Google Docs, and Copilot), and the original Arabic transcripts were manually reviewed to ensure translation validity and reliability. The data were analyzed using qualitative content analysis with MAXQDA Pro 2.0 software, applying the deductive coding approach based on the pre-set categories.
Results
The participants generally expressed positive perceptions of learning Turkish. AI supported language acquisition (e.g., serving as a personal teacher), while immersion remained the key factor in language learning. Although the students addressed several challenges pertaining to language acquisition (e.g., Turkish vowel pronunciation and suffixes), AI tools assisted them in overcoming some of these issues. ChatGPT and Gemini were the most suggested AI tools among the participants, along with practicing sentence structure before real-life conversations.
Conclusion
This study highlights the role of digital assistants in Turkish language acquisition, particularly in improving learners’ proficiency and confidence, despite limitations related to sample type and context. It also provides practical implications and future suggestions for integrating AI into Turkish language learning.
