Page 604 - ISC PROCEEDINGS 21.4

agreement process between the researcher and one of the experienced ESL teachers
which led to the agreement rate of 92%. This consistency was high and this fact
reinforced the credibility of the pronunciation scoring process.
5.5. Procedure
The intervention plan was done within 8 weeks. The start of the research was the
vocabulary and pronunciation pre-test which allowed students to determine their base
level of proficiency. After this, AI-based learning tools became part of the classroom
learning experience and self-study, whereby students could practice their vocabulary
individually and undergo pronunciation training. During the intervention period, the
observations in the classrooms were performed to trace the interactions and
participation of students. The post-tests, questionnaires and final classroom observation
activities were also carried out at the end of the 8-week period, which allowed the
researcher to compare the progress and gather the data based on perceptions.
5.6. Data analysis
The obtained quantitative data were processed with descriptive statistics to
summarize the performance of learners and with the help of paired-sample t-tests to find
out whether statistically significant differences between the scores were present
between pre -test and post-test. The level of significance was determined as p < 0.05 and
Cohen d was used to calculate the effect sizes in order to determine the size of the
intervention effect. In the meantime, qualitative data collected by questionnaires and
classroom observations had been analyzed based on thematic analysis, which made it
possible to identify repetitive themes connected with the engagement, autonomy, and
perceived effectiveness, as well as to interpret them.
5.7. Ethical considerations
The ethical issue was taken into consideration during the research. The participants
were notified of the aim of the study and the fact that they had the freedom to pull out at
any time without penalty. Any information that had been gathered in the research was
confidential and was used purely as an academic research. Anonymization of the
participants was done to ensure privacy and protect their personal data.
6. Results and discussion
6.1. Vocabulary acquisition
The vocabulary tests were analyzed quantitatively, and it was possible to notice that
the vocabulary knowledge in the participants had improved significantly due to the use of
AI-based learning tools. The pre-test results showed that a large number of the learners
possessed a small range vocabulary, and this influenced their proficiency to comprehend
and generate English. Nevertheless, the post-test scores showed that there was a
significant growth in the vocabulary scores following the intervention period. The
enhancement in the learning of vocabulary is credible to a number of characteristics of AI-
based learning systems. To begin with, the AI-based vocabulary learning environments
offered contextualized examples as well as multimedia tools and formative exercises,
which allowed learners to comprehend and remember new words better. Previous
studies have indicated that contextualized learning conditions have been effective in
memory of vocabulary since learners experience words in context as opposed to
memorizing word lists in isolation (Nation, 2013). Second, AI technologies provided
learners with individual learning opportunities, which meant that they were able to
concentrate on the vocabulary items that needed more drilling. These feedback systems
measured the performance of the learners and automated the level of difficulty of

603