AI in Language Education · Principle II: Responsible
Guideline 4: Foster environmental
sustainability
sustainability
AI has immense potential to transform language education — but also introduces social and environmental responsibilities. Sustainable integration requires calibrated, purposeful use.
1
The environmental footprint of AI
Hidden costs of AI infrastructure
Carbon emissions
Training large language models generates significant greenhouse gas emissions
Water consumption
Data centres require advanced cooling systems, leading to intensive water use
Energy demands
Servers, computers, and data centre infrastructure consume vast amounts of energy
2
Relative energy costs of common tasks
Energy use comparison (approximate)
AI image generation
Very high
AI text generation
Moderate
Web search
Low
Conventional search
Very low
Generating one AI image uses roughly the same energy as fully charging a smartphone. Web searches use 10–30× less energy than AI-generated content (De Vitries, 2023).
3
Sustainable classroom practices
Use calibrated prompts
Design prompts and workflows to maximise the quality of responses while minimising unnecessary computation and energy use.
Evaluate when to use AI
Ask: is AI truly needed here? For visuals, prefer open image banks. For information seeking, conventional web search is far more efficient.
Use freely accessible content
Reusing existing digital content reduces environmental load. Open-access libraries and educational repositories are more sustainable than generating new material.
Demand vendor transparency
Encourage AI providers to disclose environmental metrics — carbon and water use per operation — so educators can weigh the often invisible costs of AI deployment.
Did you know?
Many advanced LLMs like GPT-4 and Claude 3 Sonnet have been designed to reduce bias — yet they still exhibit implicit biases: associating negative language more often with Black individuals, linking women with humanities over science, and favouring men in leadership contexts (Maslej et al., 2025).
4
Competence checklist
COMPETENCE 16
Recognise social & environmental consequences
- How do the tools I use affect people and ecosystems beyond my classroom?
- What invisible costs (energy, data, labour) accompany digital teaching actions?
COMPETENCE 17
Make sustainable pedagogical choices
- Do I consider low-impact alternatives when planning lessons?
- How do I teach learners to use AI efficiently and responsibly?
COMPETENCE 18
Contribute to collective responsibility for sustainable AI use
- How do I share examples of environmentally responsible practice with colleagues?
- What institutional policies or collaborations could reduce the environmental and social footprint of AI in education?
