Improving Student Writing With AI
Willy A Renandya, 16 March 2025
Second Language Acquisition (SLA) theory provides a strong foundation for integrating Artificial Intelligence (AI) into EFL writing instruction, ensuring that technology supports key learning processes. Research shows that effective language acquisition requires exposure to comprehensible input, opportunities for meaningful language production, interaction, attention to linguistic features, and progressive skill development.
AI can enhance writing development by providing level appropriate models, encouraging engagement with language through interactive exercises, and making linguistic patterns more noticeable through targeted feedback. Chatbots and collaborative platforms simulate authentic interaction, while adaptive writing assistants support the transition from explicit learning to fluent application through structured practice.
By aligning AI-driven instruction with SLA principles, language teachers can use technology not just to automate tasks but to foster deeper and more effective language acquisition, helping EFL students progress from basic text construction to producing complex, coherent, and accurate texts.
The Role of AI in Second Language Acquisition
AI-powered tools are designed to provide personalized learning experiences, immediate feedback, and scaffolded support, all of which are essential for language learners. These tools align with key SLA principles:
- Input Hypothesis: Learners need comprehensible input to internalize language structures.
- Output Hypothesis: Learners must produce language to develop fluency and accuracy.
- Interaction Hypothesis: Interaction with others or tools helps learners refine their skills.
- Noticing Hypothesis: Learners improve by noticing and correcting their errors.
- Skill Acquisition Hypothesis: Skills development requires extensive and supervised practice.
By thoughtfully using AI, teachers can create a learning environment that supports these principles, helping students progress from A2 to B1 in writing.
AI Tools and Strategies for Writing Development
1. AI-Generated Writing Models
AI tools like ChatGPT or DeepSeek can generate sample essays, stories, or paragraphs based on specific prompts. These models serve as comprehensible input, allowing students to analyze well-structured texts and learn how to organize their own writing. For example, a student struggling with essay structure can study an AI-generated model and then apply similar techniques in their work. This approach helps students internalize grammar, vocabulary, and text organization, which are essential for progressing to B1-level writing.
2. Interactive Writing Prompts with Scaffolding
AI tools such as Quillbot provide step-by-step writing prompts with hints and suggestions. These tools guide students through the writing process, offering scaffolding that helps them produce more complex and accurate texts. For instance, a student writing a narrative might receive prompts like, “Describe the setting using sensory details,” helping them expand their ideas and improve their descriptive writing skills.
3. AI-Powered text Expansion
Tools like LanguageTool or Grammarly suggest ways to expand simple sentences into more complex ones. This encourages students to practice pushed output, moving from basic sentences (e.g., “I like coffee”) to more sophisticated structures (e.g., “I enjoy drinking coffee in the morning because it helps me feel energized”). By practicing sentence expansion, students develop the ability to write longer, more detailed texts, a key skill for B1-level writing.
4. AI-Enhanced Peer Feedback
Platforms like Peergrade or Eli Review use AI to facilitate peer review by guiding students to give constructive feedback. This process fosters meaningful interaction, as students learn from each other’s writing and refine their own skills. For example, a student might receive feedback like, “Your introduction is clear, but you could add more details to support your main idea.” This collaborative approach helps students develop critical thinking and writing skills.
5. AI-Powered Repeated Writing Practice with Feedback
AI tools like Write & Improve or Criterion provide students with repeated writing tasks and instant feedback on their work. This aligns with the skill acquisition hypothesis, which emphasizes that skills development requires extensive and supervised practice. For example, a student might write multiple drafts of an essay, with AI providing feedback on grammar, vocabulary, and coherence each time. This iterative process helps students refine their writing skills through consistent practice and targeted improvement.
6. AI-Powered Dialogue Writing
AI chatbots such Gemini or DeepSeek simulate conversations, and students write dialogues based on these interactions. This activity encourages interactive writing, helping students develop fluency and coherence. For example, a student might write a dialogue between two characters, practicing natural language use and conversational tone. This type of practice is particularly useful for developing writing skills in informal contexts.
7. AI-Driven Error Analysis and Focused Practice
Tools such as Grammarly analyze students’ writing errors and provide targeted exercises to address them. This aligns with the noticing hypothesis, as students become aware of their mistakes and learn to correct them. For example, a student who frequently makes tense errors might receive exercises focused on verb conjugation. By addressing specific weaknesses, students can improve their accuracy and confidence in writing.
8. AI-Generated Story Starters
AI tools like Plot Generator or Storybird provide creative story starters or prompts. These tools encourage creative writing, helping students experiment with new vocabulary and structures. For instance, a prompt like “Write about a character who discovers a hidden world” can spark creativity and help students practice narrative writing. This type of activity is particularly engaging for students and helps them develop their expressive writing skills.
9. AI-Powered Genre-Specific Writing Practice
AI tools can generate examples of different text types (e.g., emails, reports, narratives) and guide students in writing their own. This provides genre-specific input, helping students adapt their writing style to different contexts. For example, a student might analyze an AI-generated formal email and then compose their own. By practicing different genres, students become more versatile writers, a key requirement for B1-level proficiency.
10. AI-Enhanced Collaborative Writing
Platforms like Google Docs with AI add-ons students to collaborate on writing tasks while receiving AI-generated suggestions. This fosters collaborative writing, as students work together to produce texts and receive real-time feedback. For example, a group of students might co-write a report, with AI providing suggestions on grammar and style. This collaborative approach promotes interaction and shared learning, helping students refine their writing skills.
Conclusion
AI-powered tools offer transformative opportunities for EFL students to develop their writing skills. By providing comprehensible input, encouraging pushed output, fostering interaction, and helping students notice and correct errors, these tools align with key principles of second language acquisition. Importantly, the skill acquisition hypothesis emphasizes that skills development requires extensive and supervised practice. AI tools provide the scaffolding and feedback necessary to make this practice effective, but it is the students’ commitment to regular writing that ultimately drives their progress.
Integrating these tools into the classroom can create a more engaging, personalized, and effective learning experience for students. By combining the power of AI with the principles of SLA, we can equip our students with the skills they need to succeed in their language learning journey.
Click images below for more interesting reads
I wonder how how the use of AI examples can be viewed in relation to the aim of having students examine and imitate “authentic” language. There used to be a lot of criticism of textbooks with “inauthentic” language, texts just created by textbook writers, but not reflecting actual usage. I wonder if although AI based on large language models is supposedly using authentic texts, can their creations be considered authentic? Or do we think it is just good enough?