Glossary

A
  • Accessibility — The practice of designing learning materials, tools, and activities so that people with different abilities and needs can use them.
  • Adaptive learning — A learning approach in which content, difficulty, pace, or feedback changes in response to a learner’s performance or needs.
  • AI literacy — The knowledge and skills needed to understand, use, evaluate, and discuss AI systems critically and responsibly.
  • AI model — A computational system trained to identify patterns in data and use them to produce predictions, classifications, recommendations, or generated content.
  • AI output — Any content or result produced by an AI system, such as text, an image, a translation, a score, or a recommendation.
  • AI system — A technology-based system that uses data and computational methods to generate outputs that can influence decisions, learning activities, or real-world actions.
  • AI-assisted assessment — Assessment in which AI supports tasks such as generating questions, analysing responses, or drafting feedback, while a human remains responsible for the final judgement.
  • Algorithm — A defined set of rules or computational steps used to process information and produce a result.
  • Anthropomorphism — Treating an AI system as if it had human feelings, intentions, understanding, or consciousness.
  • Artificial intelligence (AI) — A broad term for computer systems that perform tasks commonly associated with human intelligence, such as recognising patterns, understanding language, making predictions, or generating content.
  • Assessment validity — The extent to which an assessment actually measures the knowledge or skills it is intended to measure.
  • Attribution — Clearly identifying the source or creator of material, including content produced or substantially modified with AI.
  • Augmentation — The use of AI to extend or support human capabilities without replacing human responsibility and judgement.
  • Automation — The use of technology to carry out a task with limited or no direct human action.
  • Autonomy — In Self-Determination Theory, the feeling that one has meaningful choice and ownership of one’s actions; in learning, it also refers to a learner’s ability to manage and direct their own learning.
  • Bias — A systematic tendency that can lead an AI system, dataset, decision, or teaching practice to favour or disadvantage particular people, groups, languages, or viewpoints.
  • Chatbot — A computer application designed to interact with users through written or spoken conversation.
  • Cognitive load — The amount of mental effort used in working memory while completing a task or learning something new.
  • Cognitive outsourcing — Transferring thinking tasks, such as remembering, writing, analysing, or deciding, to a tool; excessive outsourcing can reduce opportunities to practise and develop those skills.
  • Competence — In Self-Determination Theory, the feeling of being capable and able to make progress through effective action.
  • Context — The background information, instructions, examples, and constraints that help a person or an AI system interpret a task appropriately.
  • Copyright — The legal protection given to original creative works, which affects how texts, images, recordings, and other materials may be copied, adapted, shared, or used to train and operate AI systems.
  • Critical AI literacy — The ability to use AI while questioning how it works, whose interests it serves, what assumptions it contains, and what social, ethical, or educational effects it may have.
  • Critical thinking — The disciplined process of analysing information, checking evidence and assumptions, comparing alternatives, and forming a reasoned judgement.
  • Cultural bias — A tendency to represent one culture’s norms, language varieties, or perspectives as standard while overlooking or misrepresenting others.
  • Data — Recorded information that can be collected, analysed, or used by a computer system, including text, images, audio, scores, and interaction records.
  • Data privacy — The principles and practices governing how information about people is collected, accessed, used, stored, and shared.
  • Data protection — The legal and organisational measures used to keep personal data safe and ensure that it is processed fairly and lawfully.
  • Deep learning — A type of machine learning that uses multi-layered neural networks to learn complex patterns from large amounts of data.
  • Design thinking — A human-centred process for understanding needs, defining problems, generating ideas, building prototypes, and improving solutions through testing.
  • Differentiation — Adapting teaching, materials, support, or expected outcomes to respond to learners’ different readiness levels, interests, and needs.
  • Digital divide — Unequal access to devices, connectivity, digital skills, support, or high-quality digital services.
  • Disclosure — An explicit statement explaining whether and how AI was used in producing a piece of work or making a decision.
  • Environmental impact — The effects that developing and using AI can have on energy use, water consumption, hardware production, emissions, and electronic waste.
  • Ethical AI — The development and use of AI in ways that respect human rights, fairness, privacy, safety, transparency, accountability, and human agency; often also called responsible AI.
  • Evidence-informed practice — Professional decision-making that combines relevant research evidence with practitioner expertise, learner needs, and the local teaching context.
  • Fairness — The aim of avoiding unjustified advantages or disadvantages for particular individuals or groups in AI-supported processes and outcomes.
  • Feedback — Information that helps a learner understand current performance and decide how to improve; AI-generated feedback should be checked for accuracy, relevance, tone, and pedagogical value.
  • Foreign language anxiety — Worry, tension, or fear associated with using or learning a language, especially when speaking, being evaluated, or making mistakes in front of others.
  • Formative assessment — The collection and use of evidence during learning to identify needs, provide feedback, and guide the next steps in teaching and learning.
  • Futures thinking — A structured way of exploring how the future could develop, rather than trying to predict one certain future.

Generative AI — AI that produces new content, such as text, images, audio, video, or code, in response to instructions or examples.

  • Hallucination — An AI-generated statement or detail that appears plausible but is false, unsupported, or invented.
  • Higher-order thinking skills — Complex thinking processes such as analysing, evaluating, synthesising, solving problems, and creating, rather than simply recalling information.
  • Human-centred pedagogy — Teaching that puts learners’ development, relationships, needs, agency, and well-being ahead of the capabilities or convenience of a technology.
  • Human-in-the-loop — A process in which a person actively reviews, corrects, approves, or takes responsibility for an AI system’s input, output, or decision.
  • Inclusion — The active removal of barriers so that diverse learners can participate meaningfully, feel respected, and have equitable opportunities to succeed.
  • Inference — The stage at which a trained AI model uses what it has learned to generate an output for new input.
  • Intellectual property — Creations of the mind and the legal rights connected to them, including copyright, trademarks, patents, and related rights.
  • Intercultural competence — The ability to communicate and interact appropriately and thoughtfully with people who have different cultural identities, experiences, and perspectives.
  • Large language model (LLM) — A type of AI model trained on very large collections of language data to predict and generate sequences of text.
  • Learner autonomy — A learner’s capacity and willingness to take increasing responsibility for goals, strategies, monitoring, and reflection in their own learning.
  • Learning outcome — A clear statement of what learners should know, understand, or be able to do after a learning activity or course.
  • Licensing — The rules or permission that specify how a work, dataset, software tool, or other resource may be used, changed, and shared.
  • Machine learning — A branch of AI in which computer systems learn patterns from data instead of relying only on rules written explicitly by programmers.
  • Model training — The process of adjusting an AI model using data so that it learns patterns useful for a particular type of task.
  • Multimodal AI — AI that can process or generate more than one type of content, such as text, images, audio, or video.
  • Natural language processing (NLP) — The field of AI concerned with enabling computers to analyse, interpret, and generate human language.
  • Novice learner — A learner with limited prior knowledge or experience in a domain who usually benefits from more explicit guidance and scaffolding than an expert learner.

Open-AI assessment — An assessment design in which learners are explicitly permitted to use AI under stated conditions and are assessed on both the outcome and the quality of their process, judgement, or disclosure.

  • Personal data — Information relating to an identified or identifiable person, such as a name, email address, voice recording, image, location, or learner profile.
  • Personalisation — The adaptation of learning content, pathways, examples, pace, or feedback to an individual learner’s goals, preferences, or performance.
  • Plausible future — A future that could reasonably occur given current knowledge and credible developments, even if it is not the most likely outcome.
  • Predictive AI — AI used to estimate a likely category, score, behaviour, or future outcome from patterns in existing data.
  • Professional foresight — The disciplined exploration of possible future changes to support better decisions and preparation in professional practice.
  • Prompt — The instruction, question, data, or other input given to a generative AI system to guide its response.
  • Prompt engineering — The deliberate design, testing, and refinement of prompts to obtain outputs that are more relevant, reliable, and suitable for a particular purpose.
  • Relatedness — In Self-Determination Theory, the feeling of being connected to, cared for by, and significant to other people.
  • Retrieval practice — Strengthening learning by actively recalling information from memory instead of only rereading or reviewing it.
  • Risk-based approach — A way of selecting safeguards according to the likelihood and seriousness of potential harm; higher-risk uses require stronger oversight and controls.
  • Scaffolding — Temporary guidance or support that helps a learner complete a task they could not yet perform independently, with the support gradually reduced as competence grows.
  • Scenario — A coherent description of a possible future situation used to explore consequences, choices, and responses.
  • Scenario matrix — A framework that combines two important and uncertain drivers of change to create four contrasting future scenarios.
  • Scenario planning — A structured method for developing and examining several plausible futures in order to test assumptions and make more resilient decisions.
  • Self-Determination Theory (SDT) — A theory of motivation proposing that autonomy, competence, and relatedness support engagement, well-being, and more self-directed motivation.
  • Signal of change — An early and often small piece of evidence that may indicate an emerging development with future importance.
  • Spaced practice — Revisiting learning across several sessions separated by time, rather than concentrating all practice in one session.
  • Stereotype — An oversimplified generalisation about a group that can shape data, AI outputs, expectations, and decisions unfairly.
  • Stochastic output — An output influenced by probability, meaning that the same prompt can produce different responses on different occasions.
  • Sustainability — Meeting present educational and social needs while considering long-term environmental, economic, and human consequences.
  • Token — A small unit of text processed by a language model; a token may be a whole word, part of a word, punctuation, or another character sequence.
  • Training data — The examples used to teach an AI model patterns; their quality, coverage, and biases influence the model’s behaviour.
  • Transparency — Making relevant information about an AI system and its use understandable, including its purpose, limitations, data practices, and role in a decision or learning task.
  • Trend — A sustained direction of change observed over time, distinct from a brief event or an early signal.
  • Uncertainty — The condition of not knowing exactly what is true or what will happen; responsible AI use and futures thinking make uncertainty visible rather than hiding it.
  • Universal Design for Learning (UDL) — A framework for designing flexible learning environments that offer multiple ways to engage, access information, and demonstrate learning.
  • Zone of Proximal Development (ZPD) — The range between what a learner can do independently and what they can achieve with appropriate guidance or collaboration.