514
Muhammad Hafidz / JEs .
JES .
E-ISSN 3031-4232
JES
Journal of Educational Studies https://ejournal.
org/index.
php/JEs Enhancing the Use of ChatGPT for Self-Talk Practice: Integrating Gamification and Artificial Intelligence into Arabic Language Education Muhammad Hafidz1.
Amrini Shofiyani2.
Muhammad Yasin Fatchul Barry3 Universitas KH.
Abdul Chalim Mojokerto.
Indonesia Universitas KH.
Wahab Hasbullah.
Indonesia UIN Maulana Malik Ibrahim Malang.
Indonesia Corresponding Author.
E-mail: hafidzikhac@gmail.
Abstract Conversation skills often pose a major obstacle for self-directed learners of Arabic due to the limited availability of conversation partners and authentic practice environments.
The emergence of artificial intelligence, such as ChatGPT, presents new opportunities.
however, its use for conversational practice remains suboptimal due to the absence of structured pedagogical approaches, including gamification elements that could enhance The main research question is: How can the use of ChatGPT as a virtual conversation partner for self-directed conversational practice be improved by integrating principles of gamification and artificial intelligence to effectively enhance Arabic speaking skills?.
This research aims to: .
design a model for self-directed conversational practice based on ChatGPT and integrated with gamification elements .
oints, levels, challenges, immediate feedbac.
test the effectiveness of this model in improving speaking fluency, vocabulary mastery, and learner self-confidence.
identify the strengths and limitations of ChatGPT as a conversational agent in the context of Arabic language teaching.
The research adopts a design-based research methodology and consists of three phases: needs analysis, model development, and a limited-scale Participants are 30 self-directed learners of Arabic at the basic and intermediate levels.
Instruments include oral tests .
re- and post-), an observation form for interactions with ChatGPT, a motivation questionnaire, and semi-structured Quantitative data will be analyzed using paired-samples t-tests, while qualitative data will be analyzed using thematic analysis.
The research is expected to provide: .
a structured protocol for self-directed conversational practice using ChatGPT with gamified scenarios .
, daily tasks, conversation challenges, a point syste.
empirical evidence of significant improvement in speaking fluency .
ith a minimum Muhammad Hafidz / JEs .
513-522 target of 25%) and learning motivation compared to traditional methods.
practical recommendations on how to formulate effective prompts for ChatGPT to generate responses appropriate to learners' proficiency levels and to support topic-based conversational practice.
Optimizing the use of ChatGPT for self-directed conversational practice requires pedagogical design that not only relies on AI capabilities alone but also incorporates gamification principles to maintain engagement and sustainability in The findings of this research are expected to serve as a reference for developers of Arabic teaching materials and for self-directed learners in using AI effectively.
Further research is recommended to test this model over a longer period and with a broader Keywords: self-directed learning, gamification, artificial intelligence.
Arabic language Introduction Background of the Problem Arabic language teaching in the digital age faces both challenges and new opportunities One of the most difficult skills for self-directed learners to acquire is conversation .
, due to the limited opportunities for interaction with native speakers and the weakness of practice environments that support spontaneous dialogue (Krashen, 1.
In language learning theory, productive skills such as speaking require repeated practice with immediate feedback (Swain, 2.
The emergence of generative artificial intelligence, particularly ChatGPT (OpenAI, 2.
, represents a paradigm shift in language education.
ChatGPT can serve as a virtual conversation partner that can be engaged with at any time, on various topics, while providing responses that adapt to the user's proficiency level.
However, the reality indicates that most self-directed learners have not utilized ChatGPT optimally for conversational practice.
they use it haphazardly without a clear pedagogical strategy, rendering its effectiveness low (Kohnke et al.
, 2.
Furthermore, the issue of motivation constitutes a critical factor.
Self-directed speaking practice is often abandoned quickly because it is boring and lacks challenge.
This is where gamificationAii.
, the application of game elements such as points, levels, badges, and challenges in non-game contexts (Deterding et al.
, 2.
Aioffers a solution.
Gamification has proven effective in increasing the engagement and persistence of language learners (Zou & Xie, 2.
Muhammad Hafidz / JEs .
513-522 Therefore, this research poses the following main question: How can the use of ChatGPT as a virtual conversation partner for self-directed conversational practice be improved by integrating the principles of gamification and artificial intelligence to effectively enhance Arabic speaking Research Objectives Based on the problem formulation above, this research aims to :
Design a model for self-directed conversational practice based on ChatGPT and integrated with gamification elements .
oints, levels, challenges, immediate feedbac.
Test the effectiveness of this model in improving speaking fluency, vocabulary mastery, and learner self-confidence.
Identify the strengths and limitations of ChatGPT as a conversational agent in the context of Arabic language teaching.
Benefits of the Research Theoretically, this research enriches studies related to the integration of artificial intelligence into Arabic language teaching, particularly with regard to productive skills.
Practically, the research findings provide tangible evidence for self-directed learners.
Arabic language teachers, and educational material developers on how to use ChatGPT in a structured and engaging manner.
Previous Studies and Theoretical Framework The Concept of Conversation in Arabic Language Teaching Conversation, linguistically, is dialogue.
In Arabic language pedagogy, conversation is a productive oral skill that requires the learner to form sentences spontaneously within a communicative context (Effendy, 2.
Conversation differs from reading and writing in that it requires speed of thinking, mastery of active vocabulary, and the courage to make mistakes.
The Arabic language curricula in many educational institutions still focus on grammatical aspects, leaving very limited time for conversational practice (Wekke, 2.
ChatGPT in Language Education ChatGPT It is a large language model trained on a vast amount of text.
It can generate human-like responses, understand conversational context, correct errors, and adjust the level of language difficulty (OpenAI, 2.
A study by Kohnke et al.
showed that ChatGPT can be an effective partner for practicing writing and speaking in English.
However, research Muhammad Hafidz / JEs .
513-522 on the use of ChatGPT for Arabic, particularly for conversation, remains very limited.
ChatGPT in Language Education Gamification is not a complete game, but rather the use of game elements .
oints, leaderboards, badges, levels, challenges, immediate feedbac.
within educational activities.
The meta-analysis conducted by Bai and colleagues .
concluded that gamification significantly improves intrinsic motivation, cognitive engagement, and language learning In speaking skills, gamification helps reduce anxiety because the atmosphere becomes more playful and positively competitive (Hung, 2.
Integration of Artificial Intelligence and Gamification The integration of generative artificial intelligence with gamification constitutes a new field known as adaptive gamification.
AI can dynamically adjust the level of challenge based on user performance, provide personalized feedback, and generate countless dialogic scenarios (Kapp & Defelice, 2.
This combination is highly promising for self-directed conversational practice because it overcomes the two main obstacles: boredom and the lack of an interactive Method Approach and Design This research employs the Design-Based Research (DBR) approach developed by Collins et .
DBR was selected because the aim is to develop and test a pedagogical intervention in a real-world context.
The DBR design consists of three iterative phases: .
needs analysis and problem identification.
model development.
limited-scale experiment and evaluation.
Approach and Design The participants are 30 self-directed learners of Arabic, meeting the following criteria: .
aged between 18 and 35 years.
basic to intermediate proficiency level .
ccording to the Common European Framework of Reference for Languages.
A2AeB.
have not previously used ChatGPT in a structured manner for conversational practice.
have access to the internet.
Participants were recruited voluntarily through online Arabic language learning communities .
Muhammad Hafidz / JEs .
513-522 Research Procedures Phase One: Needs Analysis A Conduct a preliminary survey .
uestionnaires and interview.
with 50 self-directed learners to identify difficulties in conversation and their experiences with artificial intelligence .
Formulate the initial design criteria for the model.
Phase Two: Model Development A Design a conversation protocol with ChatGPT that includes: initial prompts .
efining ChatGPT's role as a conversation partner, language level, topi.
, gamification scenarios .
oint system, levels from beginner to expert, daily tasks, timed challenge.
Prepare a written guide and an instructional video for participants.
Validate the model by three experts .
n Arabic language teaching, educational technology, and Phase Three: Limited-Scale Experiment Pre-test: Measure initial speaking ability .
luency, vocabulary, confidenc.
through an oral test with an examiner.
Intervention: Participants use the model for 4 weeks .
Ae30 minute sessions daily, 5 days per A Post-test: An oral test similar to the pre-test.
Administration of a motivation questionnaire and conduct of semi-structured interviews.
Research Instruments TOOL Oral test .
re-pos.
Observation form for ChatGPT Motivation .
dapted from IMI) Semi-structured interview guide FUNCTION Measure fluency .
ords per minut.
, grammatical accuracy, and active vocabulary mastery Record types of prompts.
ChatGPT's responses, and interaction Measure intrinsic motivation .
nterest, effort, perceived valu.
on a 1Ae5 scale Investigate participants' perceptions regarding the advantages and limitations of ChatGPT Data Analysis Quantitative data: Analysis of pre-test and post-test results using a paired-samples t-test .
f the data are normally distribute.
or the Wilcoxon signed-rank test .
f not normally distribute.
at a significance level of = 0.
Descriptive analysis of motivational improvement will also be Muhammad Hafidz / JEs .
513-522 Qualitative data: Analysis of interview transcripts using thematic analysis (Braun & Clarke, 2.
coding, theme clustering, and interpretation.
Results Expected Results Based on the above research design, the expected results include three main components:
Protocol for a Conversational Practice Model Integrating ChatGPT and Gamification This protocol includes:
A Basic rules: Defining ChatGPT's identity:
"You are a friendly Arabic conversation partner.
Speak at A2 level.
Use simple sentences.
Give 1 positive feedback and 1 correction per response".
Arabic translation provided in the original AOAO O A A CI IUE uOO OA.
AIUE OA
A I A.
A2 A IOOA.
A"I OE I OO EOA
Ae A A Point system: Participants record points themselves .
r using a spreadshee.
according to: task completion .
, use of 3 new vocabulary items .
, answering without long pauses .
A Levels:
Oe Level 1 (A.
: Everyday topics .
amily, food, hobbie.
Oe Level 2 (B.
: General topics .
ravel, education, technolog.
Oe Level 3 (B.
: Abstract topics .
pinions, culture, social issue.
A Daily tasks .
Oe Day 1: Introducing oneself .
question-and-answer round.
Oe Day 3: Ordering food in a restaurant .
Oe Day 7: Narrating yesterday's activities .
ast tens.
A Additional challenges: Recording a conversation with ChatGPT for 3 minutes without interruption and sending it to the learning group for feedback.
Empirical evidence of improvement in speaking skills Muhammad Hafidz / JEs .
513-522 A From a preliminary experiment .
onducted with 10 participants, unpublishe.
, initial indicators emerged:
Oe Increase in fluency .
ords/minut.
from 25 words to 38 words .
mprovement rate of 52%).
Oe Increase in active vocabulary from 120 words to 210 words.
Oe Increase in confidence level .
Ae5 scal.
1 to 3.
A In the main research, the minimum target is a 25% improvement in fluency .
onsistent with findings from AI speaking studies.
Li et al.
, 2.
Recommendations on effective prompts with ChatGPT Based on the analysis of interactions, optimal prompt patterns will be formulated, such A For beginners (A.
"ChatGPT, now you are a patient conversation partner.
I am a beginner in Arabic.
Talk to me about 'at the market' using short sentences.
Correct my mistakes gently.
A For intermediate learners (B.
"ChatGPT, let us have a light discussion.
Topic: Is online learning better than face-to-face learning? Give me 30 seconds to respond to each question.
A A prompt specifically for feedback:
AuPlease evaluate my following Arabic response: .
Give it a score from 1 to 10 for grammar and vocabulary.
Identify two main errors and correct themAy.
Advantages and Limitations of ChatGPT Based on the expected qualitative data, the advantages of ChatGPT include:
Available 24 hours a day, without shame or fear of judgment.
The same topic can be repeated many times.
Quick and varied responses.
Capable of adapting to the language level .
As for the potential limitations:
Not always accurate in correcting Arabic grammar rules .
ecause Arabic training data may be insufficien.
Lacks non-verbal expressions .
one of voice, gesture.
that are important in Muhammad Hafidz / JEs .
Requires an internet connection and technical understanding .
o write prompt.
Discussion The integration of ChatGPT and gamification offers a solution to two main problems in selfdirected conversation practice: .
the absence of a live partner, and .
the lack of sustained ChatGPT flexibly replaces the native speaker, while gamification transforms the practice routine into a game-like activity with goals and rewards.
This aligns with SelfDetermination Theory (Deci & Ryan, 2.
, which states that autonomy .
ontrol over time and topi.
, competence .
rogress through level.
, and relatedness .
ven if virtua.
are keys to However, significant challenges remain.
ChatGPT was not specifically designed to be an Arabic teacher.
it does not always understand the distinctions between Modern Standard Arabic and colloquial dialects, and it may produce rigid or unnatural sentences.
Therefore, the researcher recommends in this model that learners use other resources .
ictionaries, grammar book.
for It is also preferable that the gamification aspect does not focus excessively on external competition .
, leaderboard.
that may cause anxiety.
rather, it should focus on personal progress ("leveling up").
This research also opens the door for subsequent developments, such as comparing the ChatGPT model with a conversational AI specifically designed for Arabic .
Lumi.
, or integrating speech-to-text technology to enable ChatGPT to respond directly to spoken input .
urrently ChatGPT is text-based.
the learner must typ.
Conclusion This research concludes that optimizing the use of ChatGPT for self-directed conversational practice does not suffice with random "speaking" but rather requires systematic pedagogical design.
The proposed model integrates: .
a conversation protocol with clear .
gamification elements .
oints, levels, task.
a feedback mechanism.
The expected results indicate noticeable improvement in fluency, vocabulary, and motivation.
However.
ChatGPT's limitations regarding the accuracy of Arabic grammar .
specially for Muhammad Hafidz / JEs .
513-522 advanced level.
and the absence of a voice feature remain important observations.
For self-directed learners: Use the prompt guide and point system developed in this study.
Do not rely solely on ChatGPT.
integrate it with self-recording of your voice or practice with peers.
For educational media developers: Create a dedicated interface .
ebsite/applicatio.
that simplifies prompt writing, stores point records, and adds voice support .
peech recognition and text-to-speec.
For future researchers: Conduct longer-term studies .
with a control group .
earners using ChatGPT without gamification versus a group using gamificatio.
Also, study the effectiveness of the model for advanced learners (B2AeC.
For educational institutions: Consider integrating ChatGPT gamification as a selfdirected supplementary assignment to regular classroom instruction, especially to address the need for conversation practice, which is often limited.
References