ABSTRACT

This scientific-methodological article presents a Human–AI Dialogue Model for English language teaching in which generative artificial intelligence functions as an additional language partner rather than a substitute for the teacher. The model is designed to increase meaningful speaking opportunities, develop learner agency and strengthen critical digital literacy. Its core cycle—Purpose → Prompt → Interact → Challenge → Reformulate → Reflect—moves learners from passive consumption of AI-generated content to active communication. The article introduces practical techniques including AI Role-Play, AI Error Detective, AI Debate Partner, Missing-Word Negotiation, Prompt Ladder, AI Story Builder, Feedback Loop and Dialogue Reflection. The approach combines communicative language teaching, the CEFR action-oriented perspective, mediation, formative assessment and human-centred AI use. A model lesson sequence and assessment framework demonstrate how AI can support real-world English communication while keeping the learner and teacher at the centre of the educational process.

Keywords: artificial intelligence, English language teaching, human–AI interaction, language partner, generative AI, speaking, learner agency, critical thinking, formative assessment, digital literacy.

1. INTRODUCTION

Generative artificial intelligence has changed the practical conditions of language learning. Learners can now practise conversations, request alternative expressions, simulate real-life situations and receive immediate feedback outside the limited time of the classroom. Yet access to AI does not automatically produce communicative competence. The central methodological question is therefore how to design AI interaction so that students remain active speakers, thinkers and decision-makers.

This article proposes the concept of AI as a second language partner. The word “partner” is used pedagogically: AI provides additional interaction, while the teacher remains the learning designer, facilitator and ethical guide. The learner decides what to communicate, evaluates the response and improves the original language. This is compatible with the CEFR action-oriented approach, which views learners as social agents and gives importance to interaction, mediation and purposeful language use (Council of Europe, 2020).

The approach also supports learner agency and reflection, which are central to future-oriented education (OECD, 2019). At the same time, responsible use of generative AI requires human oversight, critical evaluation, privacy awareness and meaningful educational purpose (Miao & Holmes, 2023).

2. THE CORE CONCEPT: FROM AI TOOL TO LANGUAGE PARTNER

The methodological difference is simple but significant. In a traditional AI task, a student asks a question and copies an answer. In the proposed model, the student first has a communicative intention, interacts with AI, challenges or checks the response, reformulates language and reflects on the result. AI therefore becomes a stimulus for language production rather than a replacement for it.

Three roles are intentionally separated: the teacher designs the learning environment; the student is the active communicator; AI is the responsive practice partner. This structure protects the pedagogical purpose of the lesson and makes technology serve communication.

3. HUMAN–AI DIALOGUE MODEL

The cycle is recursive: reflection becomes the starting point for the next dialogue. As learners progress, the teacher increases the complexity of roles, vocabulary, interaction rules and problem situations.

4. HIGH-IMPACT AI-BASED METHODS AND TECHNIQUES

4.1. AI ROLE-PLAY AI becomes a realistic communication partner: tourist, customer, employer, doctor, colleague or international student. The learner must adapt vocabulary, tone and politeness to the situation.

4.2. AI ERROR DETECTIVE AI generates a short text containing selected grammar, vocabulary or register problems. Students identify, explain and correct the errors, then ask AI to compare the revised version. The learner becomes the language evaluator.

4.3. AI DEBATE PARTNER AI takes an opposing position. The learner must present a claim, reason, example and counterargument. This combines speaking fluency with critical thinking.

4.4. MISSING-WORD NEGOTIATION AI does not provide an unknown word directly. It gives a definition, contrast, example or clue. Students paraphrase, guess and negotiate meaning, developing circumlocution.

4.5. PROMPT LADDER Students move through cognitive levels: Describe → Explain → Compare → Justify → Challenge → Create. The same topic becomes progressively more demanding.

4.6. AI STORY BUILDER The learner and AI build a story one turn at a time. The student must continue the narrative using target vocabulary and a required grammatical structure.

4.7. AI FEEDBACK LOOP Generate → Check → Correct → Improve. Students compare their original answer with AI suggestions but must decide which changes are actually useful.

4.8. DIALOGUE REFLECTION After speaking, students complete three prompts: “A phrase I used well…”, “A mistake I noticed…”, “A phrase I will use next time…”.

AI METHOD TOOLKIT — SLIDE-STYLE OVERVIEW

5. FORMATIVE ASSESSMENT: MEASURING HUMAN PERFORMANCE

Assessment should measure the learner’s communicative performance, not the sophistication of the AI tool. A practical rubric can evaluate five dimensions: communicative success, fluency, lexical flexibility, interaction strategy and reflection. This keeps the student’s contribution visible and discourages copying.

6. 45-MINUTE MODEL LESSON

7. RESPONSIBLE AND ETHICAL USE OF GENERATIVE AI

AI-supported language learning requires clear boundaries. Students should not share sensitive personal information, should verify important claims and should understand that AI-generated language may be inaccurate, biased or contextually inappropriate. The teacher should establish age-appropriate rules and distinguish support from authorship. UNESCO recommends a human-centred approach to generative AI in education that prioritises safety, equity, privacy and meaningful learning (Miao & Holmes, 2023).

A practical classroom rule is: “Create first, consult second, improve third.” The student produces an initial response before using AI. AI is then used for comparison, alternative wording and revision. This preserves productive struggle and learner ownership.

8. EXPECTED OUTCOMES AND PRACTICAL SIGNIFICANCE

The model increases opportunities for meaningful interaction and can support fluency, lexical retrieval, paraphrasing, negotiation of meaning and self-correction. It also allows differentiated practice: students can repeat a situation, change the level of difficulty or request a new role. Importantly, the teacher remains responsible for learning outcomes, assessment, task quality and ethical boundaries.

The expected long-term outcome is learner independence. Students should gradually become able to initiate dialogue, maintain interaction, repair communication, question information, improve their own language and reflect on progress. In this sense, AI is valuable not because it speaks more than the learner, but because it creates more reasons for the learner to speak.

9. CONCLUSION

AI can function as a second language partner when interaction is pedagogically designed. The Human–AI Dialogue Model transforms generative AI from an answer-producing tool into a structured environment for communication, challenge, reformulation and reflection. By combining communicative practice, learner agency, mediation, formative assessment and critical digital literacy, the model supports a future-oriented English classroom.

The central principle is clear: technology should not reduce the learner’s thinking or speaking. It should expand opportunities to communicate. The most successful AI-supported English lesson is therefore not the one in which AI produces the most content, but the one in which the student produces more meaningful language, makes more informed choices and becomes a more independent communicator.

REFERENCES

Council of Europe. (2020). Common European Framework of Reference for Languages: Learning, Teaching, Assessment – Companion Volume. Strasbourg: Council of Europe.

Miao, F., & Holmes, W. (2023). Guidance for Generative AI in Education and Research. Paris: UNESCO.

OECD. (2019). OECD Learning Compass 2030. Paris: OECD.

OECD. (2025). OECD Teaching Compass: Reimagining Teachers as Agents of Curriculum Change. OECD Education Policy Perspectives, No. 123. Paris: OECD Publishing.

Richards, J. C. (2006). Communicative Language Teaching Today. Cambridge: Cambridge University Press.

Қанатова Пәрмиза Қанатқызы