AI in Language Teaching

Artificial intelligence is transforming English language teaching through automated writing feedback, conversational chatbots, adaptive learning platforms, AI-generated materials, and machine translation as a learning tool. The field is evolving rapidly. Applications that were experimental in 2020 became mainstream by 2025.

Current Applications

Application	What it does	Examples
Automated Writing Evaluation (AWE)	Analyses writing for grammar, vocabulary, coherence, and task achievement	Grammarly, Write & Improve, ETS e-rater
Conversational AI / Chatbots	Provides speaking and writing practice through dialogue	ChatGPT, Replika, language-specific bots
Adaptive learning platforms	Adjusts content difficulty and pacing based on learner performance	Duolingo, ELSA Speak, Busuu
AI-generated materials	Creates exercises, reading passages, listening scripts, and test items	LLM-based content generation
Machine translation	Supports comprehension and raises language awareness	Google Translate, DeepL
Speech recognition	Provides pronunciation feedback	ELSA Speak, SpeechAce
Learning analytics	Tracks learner behaviour and predicts outcomes	LMS-integrated analytics

Research Evidence

Meta-analytic evidence (Lyu 2025) shows AI chatbots have a medium positive effect on L2 learning (g = 0.608). Key findings:

• Chatbots reduce speaking anxiety by providing a low-stakes practice environment
• Automated feedback is most effective when combined with teacher feedback, not as a replacement
• Lower-proficiency learners may benefit more from AI interaction than advanced learners
• Task design matters more than the technology itself. AI tools need purposeful integration

Opportunities

• Scalability: AI provides individualised feedback and practice that would be impossible for a single teacher to deliver at scale
• Accessibility: 24/7 availability, so learners can practise outside class time
• Personalisation: adaptive systems respond to individual learner needs, pace, and level
• Efficiency: automated assessment frees teacher time for higher-value interactions
• Data-driven insight: learning analytics can inform pedagogical decisions

Concerns and Limitations

• Accuracy: AI feedback can be unreliable, especially for pragmatic, discourse-level, and culturally specific language use
• Overreliance: learners may develop dependency on AI tools rather than building autonomous competence
• Equity: access to AI tools requires devices, internet, and often subscriptions
• Academic integrity: AI-generated text complicates assessment of genuine learner production
• Data privacy: learner data fed to AI platforms raises ethical questions
• Teacher deskilling: if AI automates feedback and materials creation, teachers may lose professional competencies
• Bias: AI models reflect training data biases, potentially reinforcing native-speaker norms and cultural assumptions

Pedagogical Principles

Effective AI integration follows the same principles as any TELL:

1. Start with learning objectives, not the technology
2. Use AI to complement, not replace, teacher expertise
3. Train learners to evaluate AI output critically
4. Maintain human connection and communicative authenticity
5. Address equity and access proactively

References

• Godwin-Jones, R. (2022). Partnering with AI: Intelligent writing assistance and instructed language learning. Language Learning & Technology, 26(2), 5–24.
• Lyu, B. (2025). Effectiveness of chatbots in improving language learning: A meta-analysis. International Journal of Applied Linguistics, 35(1).
• Kohnke, L., Moorhouse, B. L., & Zou, D. (2023). ChatGPT for language teaching and learning. RELC Journal, 54(2), 537–550.