NIR AI Governance for ELT™
AI governance for ELT, as set out in NIR AI Governance for ELT™, is an institutional guidance framework for the responsible, ethical, transparent, pedagogically sound, and sustainable use of AI in English Language Teaching.
Pedagogy First. AI Enhanced. Future Ready.
Framework status. NIR AI Governance for ELT™ is a guidance framework for voluntary self-assessment and planning. It is not an accreditation, certification, regulatory standard, or externally validated framework. It does not provide legal advice, and institutions remain responsible for compliance in their own jurisdictions.
What is AI governance for ELT?
AI governance for ELT is the institutional system of accountability, standards, policies, and review through which an English language teaching organisation keeps AI use pedagogically justified, ethical, transparent, and under human responsibility.
NIR AI Governance for ELT™ helps teachers, academic leaders, language schools, universities, and ELT organisations decide, document, and review how AI may be used in teaching, learning, assessment, administration, and communication. It applies general responsible-AI ideas to the specific conditions of language education: construct validity in assessment, CEFR-aware appropriacy, fairness across varieties of English and learner backgrounds, and the professional agency of teachers.
How AI governance for ELT works
Governance determines who decides, what is governed, who is accountable, and how oversight works. It follows six connected stages, and review returns to the principles, so governance stays a cycle rather than a one-off document.
- PrinciplesThe values and commitments that govern all AI use.
- AccountabilityWho owns governance at institutional, programme, and classroom level.
- Decision RightsWho may approve, restrict, escalate, or stop AI uses.
- Governance Domains & StandardsWhat must be established in each of the 12 domains.
- OversightHow human oversight, risk review, and assurance evidence are used.
- ReviewScheduled and event-driven review that returns to the principles.
Six concepts, kept separate
Institutions often treat these as one thing. NIR keeps them distinct so that responsibilities, evidence, and tools are not confused.
| Concept | Core question | Level | Typical output |
|---|---|---|---|
| AI Governance | Who is accountable, and under what rules and oversight is AI used? | Institution | Governance structure, standards, decision records |
| AI Literacy | What do people understand about how AI works, its limits, and its risks? | Individual and community | Foundational knowledge and critical awareness |
| AI competency | What can a teacher do with AI in professional practice, and how well? | Teacher | Demonstrable professional practice |
| AI Quality Assurance | Does AI-supported practice, in actual use, meet intended standards? | Process and practice | Review, sampling, moderation, quality evidence |
| AI Policy | How are governance decisions documented as rules and procedures? | Document | Policies and procedures through which governance requirements are implemented |
| AI Tool Evaluation | Is this specific tool fit for this specific educational purpose? | Tool and procurement | Evaluation record and adoption decision |
A well-written policy is ineffective if staff lack understanding, or if the institution never checks practice. Each concept supports the others.
Four connected layers
Governance determines the institutional requirements. Policy documents and implements them. Quality assurance checks actual practice. Evaluation assesses specific tools and use cases.
NIR AI Governance for ELT™
Defines what institutional governance must establish and oversee: who decides, what is governed, who is accountable, and how oversight works.
AI Policy Center
Shows how governance requirements are documented and implemented, through policies, templates, and implementation resources.
AI Quality Assurance for ELT
Checks actual practice against the intended standards and supplies evidence to governance decisions.
AI Evaluation & ELT Expertise
Evaluates specific AI tools and use cases against ELT criteria.
Ten principles of AI governance for ELT
The principles are conceptually aligned with widely discussed ideas in responsible AI, human-centred education, and educational ethics. They are not derived from, or endorsed by, any particular body.
- Pedagogy first. Educational purpose comes before tool selection.
- Human responsibility. AI output informs professional judgement; it does not replace it.
- Learner-centred. AI serves learners’ development, wellbeing, and agency.
- Teacher agency. Teachers can accept, adapt, or reject AI output.
- Transparency. People are appropriately informed when AI materially affects them or their educational experience.
- Integrity. AI use must not undermine the validity of learning or assessment.
- Privacy and security by design. Personal data is minimised and protected.
- Fairness and inclusion. Bias is monitored; accessibility is designed in.
- Proportionality. Controls scale with the potential for harm.
- Continuous learning. Policy and practice are reviewed against evidence.
The 12 domains of AI governance for ELT
Each domain carries institutional standards and expected evidence of practice. Adopt them in full or adapt them proportionately to your size and risk.
Governance & Leadership
Structure, roles, institutional bodies, and decision rights, integrated with existing quality structures.
Teacher AI Use & Professional Responsibility
Teachers verify AI output, keep professional authority, and receive development support.
Student AI Use & Learner Agency
Clear, accessible guidance by task type, and tasks that preserve language development.
Academic Integrity & Assessment
Declared AI-use status per assessment, valid task design, and no sole reliance on AI detection.
Data Privacy & Security
Tool and data registers, minimisation, vendor terms, and special care for sensitive learner data and data relating to minors.
Transparency & AI Disclosure
Plain-language public statements and disclosure when AI materially affects learners.
Human Oversight & Decision-Making
Impact classification, named decision-makers, and access to appropriate human review for consequential decisions.
Bias, Fairness & Accessibility
Review for effects across accents, first languages, and varieties of English; accessible tool selection.
AI Tool Evaluation & Procurement
Documented approval, evidence-based review where available, tested vendor claims, pilots, and re-evaluation on material change.
Governance Implementation & Accountability
Decision pathways, escalation, ownership across sites and units, an exceptions process, and consultation. Required policies exist, with owners and review cycles.
Governance Oversight & Risk
Institutional oversight, risk review, incident escalation, and evidence-based decisions that draw on assurance from AI Quality Assurance for ELT.
Governance Review & Improvement
Scheduled and event-driven review that returns to the principles, with owned and tracked actions.
What makes the framework specific to ELT
The framework applies governance to the conditions of language education. Six concerns run through the domains and distinguish it from general AI governance.
CEFR alignment
Level and descriptor appropriacy of AI-generated texts, tasks, and feedback is reviewed by qualified staff.
Assessment validity
AI use is judged against what a language task is meant to evidence about proficiency.
English varieties and accents
Tools are examined for how they treat accents, dialects, and World Englishes.
Learner backgrounds
First languages, multilingual repertoires, age, access, and digital confidence are considered in design and review.
Teacher agency
Teachers keep the authority to accept, adapt, or reject AI output, and are not required to use AI.
Learner agency
Learners use AI critically, with clear guidance, and without losing the effort through which language develops.
Risk and impact classification
Governance classifies each AI use by its potential impact on learners, teachers, and the institution. The level determines who may approve the use, what human oversight applies, and what assurance evidence governance requires.
Supportive uses
Typical ELT uses: drafting and brainstorming support, lesson ideas, vocabulary lists reviewed by the teacher.
Governance expectation: teacher accountability, approved tools where learner data is involved, and use recorded in the AI use register.
Learner-facing uses
Typical ELT uses: formative feedback on writing or speaking, practice and tutoring tools, learner analytics for teaching support.
Governance expectation: disclosure to learners, an accountable academic lead, and periodic assurance evidence from quality assurance.
Consequential uses
Typical ELT uses: scoring or grading support, placement, progression, certification, decisions affecting learner welfare, and sensitive data about minors.
Governance expectation: governance-group approval, named human decision-makers, access to human review, documented risk review, and regular assurance evidence.
Classification is proportionate and illustrative. Institutions define their own thresholds and reclassify a use when a tool’s function or data practices change.
AI governance for ELT: institutional maturity model
A five-level model for assessing AI governance for ELT at your institution, for self-assessment and planning. Use it with the governance checklist at each governance review. Rate each domain separately, because maturity is usually uneven. It is not a certification or a ranking.
AI use is individual and informal. No shared rules.
Leaders recognise the need. Initial guidance exists for some uses.
Policies, roles, and approval processes are documented and communicated.
Practice is monitored; indicators, audits, and incidents inform decisions.
Review is cyclical and evidence-informed; lessons are shared.
How the framework fits with NIR
NIR AI Governance for ELT™ is the institutional layer of an existing NIR structure. It does not replace any component.
AI Teacher Competency Framework™
Defines what teachers can do with AI. Governance states what institutions require; competency states what teachers can do.
AI Teacher Readiness Assessment™
Identifies readiness and development needs to inform training and implementation planning. Developmental, not disciplinary.
NIR ELT Intelligence™
Where used, its outputs are treated as decision-support under human oversight, data, and disclosure standards.
AI Quality Assurance for ELT
Puts monitoring into practice through sampling, review, and moderation, supported by AI Evaluation & ELT Expertise.
NIR Technology is held to the same standards as any other tool. Institutions should evaluate NIR products under Domain 9 with the same independence they apply to any vendor. See AI in ELT Frameworks & Intelligence for how the components connect, the NIR Ecosystem, AI Ethics in Education, and the AI Glossary for ELT.
Academic positioning and alignment
The framework is a conceptual, practice-oriented synthesis. Its contribution is contextualisation: it applies established ideas from responsible AI, educational ethics, and quality management to the specific conditions of English Language Teaching. The sources below are reference points for orientation only. They do not endorse, certify, or partner with NIR Academy, and the framework does not claim to implement them. Consult the current versions directly.
- UNESCO: artificial intelligence in education for human-centred approaches to AI in education.
- OECD AI Principles for responsible stewardship of trustworthy AI.
- NIST AI Risk Management Framework for governing and managing AI risk as a continuing activity.
- Common European Framework of Reference for Languages (Council of Europe) as a shared reference for describing language proficiency.
The framework has not been empirically validated. Each institution must determine its own legal and regulatory obligations, for example under data-protection law.
Frequently asked questions
What is NIR AI Governance for ELT™?
Is the framework a certification or accreditation?
Why does AI governance for ELT need its own framework?
How is AI governance different from AI literacy?
How is AI governance different from an AI policy?
Does the framework show institutions how to write AI policies?
Who is the framework for?
Does the framework guarantee legal compliance?
Planning AI governance for your institution?
Start with a baseline self-assessment across the 12 domains, then choose two or three priorities.
For institutions Contact NIR AcademyNIR AI Governance for ELT™ is developed by NIR Academy and is currently a working draft (Version 1.0). It is a conceptual, practice-oriented synthesis and has not been empirically validated, accredited, or endorsed. Maturity levels, indicators, and templates are illustrative and require local adaptation. NIR Academy also offers related products and services; institutions should evaluate them independently. Last reviewed: October 2026.

