Generative Artificial Intelligence and Language Teaching – Core Vocabulary

Key terms and definitions drawn from the lecture notes on Generative AI and Language Teaching, covering core technologies, pedagogical frameworks, ethical considerations, and professional development concepts.

Generative Artificial Intelligence (GenAI)

A subset of AI that creates new content (text, code, images, audio, video) in response to prompts using large data sets and deep-learning models.

Large Language Model (LLM)

An AI model trained on vast text corpora that predicts the next token in a sequence to generate human-like language.

Generative Adversarial Network (GAN)

A type of AI architecture where two neural networks (generator and discriminator) compete to create realistic synthetic data such as images or audio.

Conversational AI Chatbot

An LLM-based tool that engages in human-like dialogue, interprets prompts, and can perform tasks such as content creation, summarisation, and role play.

Visual / Audio / Video Generator

A GenAI tool that converts natural-language prompts into images, sound, or video (e.g., DALL-E, Stable Diffusion, Sora).

Embedded GenAI Function

A generative AI feature integrated into an existing digital tool (e.g., quiz builders, slide designers, LMS assistants).

Professional-GenAI-Competence (P-GenAI-C)

A five-part construct describing the knowledge and skills language teachers need to use GenAI responsibly and effectively.

GenAI Technological Proficiency

Awareness of a range of GenAI tools, their functions, affordances, and limitations.

Pedagogical Compatibility (PC)

Ability to integrate GenAI to supplement and enhance students’ language learning in line with sound pedagogy.

Teachers’ Professional Work (P-GenAI-C Aspect)

Use of GenAI for tasks outside the classroom such as grading, feedback, communication, and administration.

Ethical Use (EU)

Teachers’ awareness of risks, well-being, and ethical issues when applying GenAI tools.

Preparing Students for a GenAI World

Equipping learners with knowledge and skills to engage critically and productively with GenAI in study, work, and life.

Prompt Engineering

Crafting detailed, context-rich prompts and iteratively refining them to obtain desired GenAI outputs.

GenAI Hallucination

When an AI model produces fluent but factually incorrect or fabricated information.

Bias in GenAI

Systematic favouring or stereotyping in AI outputs resulting from imbalanced or prejudiced training data.

Transparency (Assessment)

Clarity for students and teachers regarding the purpose and criteria of an assessment.

Validity (Assessment)

The extent to which an assessment measures what it is intended to measure.

Reliability (Assessment)

Consistency of assessment results across different raters or occasions.

Self-Directed Learning (SDL)

A process in which learners take initiative for diagnosing needs, setting goals, selecting strategies, and evaluating outcomes.

AI Literacy

The ability to understand, access, prompt, corroborate, and incorporate AI tools effectively and ethically.

Understanding (AI Literacy)

Knowing an AI tool’s purpose, capabilities, limitations, and ethical considerations.

Accessing (AI Literacy)

Selecting appropriate AI platforms and features to meet specific learning objectives.

Prompting (AI Literacy)

Formulating clear, precise queries and iterating them to refine AI outputs.

Corroborating (AI Literacy)

Verifying AI-generated information with trusted external sources.

Incorporating (AI Literacy)

Ethically integrating AI outputs while preserving one’s own voice and understanding.

Micro-learning

Professional development delivered in short, focused bursts (videos, infographics, mini-courses) for just-in-time skill building.

Professional Digital Competence

Technology-related knowledge and skills specific to a profession; adapted here for GenAI contexts.

Standardisation (Language)

The dominance of certain ‘standard’ varieties of English in AI outputs, potentially marginalising other dialects.

Creative Constraint

Risk that AI-generated patterns lead learners to produce formulaic language, limiting originality.

Automatic Item Generation

Using GenAI to create test passages and questions quickly and at scale.

Temperature (LLM Setting)

A parameter controlling randomness: low values yield predictable text; high values produce diverse, creative output.

Context Window

The maximum amount of prompt and conversation history an LLM can process at once.

Training Data

Curated and formatted datasets used to teach an AI model linguistic patterns and knowledge.

Data Source

Original raw information (books, websites, transcripts) from which training data are derived.

Embedded AI Assistant

A built-in GenAI feature within software that automates sub-tasks (e.g., quiz writers in Kahoot!, co-pilots in LMS).

Iterative Design (Prompting)

A cycle of drafting, evaluating, revising, and resubmitting prompts to improve GenAI responses.

Practitioner Research

Systematic inquiry by teachers into their own practice to generate context-specific evidence.

Micro-credential

A short, focused certification or badge demonstrating mastery of a specific skill, often earned online.

Environmental Footprint of AI

Energy consumption and carbon emissions associated with training and running large language models.

Virtual Twin

An AI-generated avatar of a teacher that can interact with learners or attend meetings autonomously.