Jankova Alagjozovska, Natka (2025) Solving the Puzzle of Speech Sounds in English Language: AI Prompts for Improvement in Pronunciation. In: Artificial Intelligence (AI) in the Age of Transformation: Opportunities and Challenges, 13 May 2025, Pristina, Kosovo. (Submitted)
![- Final - 2025 Programme Conference.pdf [thumbnail of - Final - 2025 Programme Conference.pdf]](https://eprints.ugd.edu.mk/style/images/fileicons/text.png)
- Final - 2025 Programme Conference.pdf - Additional Metadata
Download (474kB)
![aab COLLEGE CONF.pdf [thumbnail of aab COLLEGE CONF.pdf]](https://eprints.ugd.edu.mk/style/images/fileicons/text.png)
aab COLLEGE CONF.pdf - Presentation
Download (537kB)
Abstract
Phonemes and allophones are speech units that help to solve the puzzle of speech sounds in English language. Each system is language specific. Hence, there is a big difference between Macedonian and English language. Very often differences are not accessible to the sense of hearing. By using modern phonetic instruments, sounds are transformed into visible records which make differences more noticeable. Additionally, AI can be also helpful in defining and distinguishing allophones and phonemes. This study will define phonemes and allophones while also providing phoneme-based and allophone-based AI prompts for pronunciation improvement. Variations in sounds occur across connected speech for various reasons, such as background noise, the speaker’s fatigue, distraction, or reluctance to repeat. Furthermore, the speaker’s linguistic background can influence these variations. Studying and improving both distinctive and non-distinctive features can help in understanding these changes, which often lead to miscommunication.
Key words: Phonemes, Allophones, AI, Pronunciation, English language, Speech.
Item Type: | Conference or Workshop Item (Paper) |
---|---|
Subjects: | Humanities > Languages and literature |
Divisions: | Faculty of Philology |
Depositing User: | Natka Alagozovska |
Date Deposited: | 14 May 2025 07:09 |
Last Modified: | 14 May 2025 07:09 |
URI: | https://eprints.ugd.edu.mk/id/eprint/35953 |