Generative AI
31 Terms
A–Z Linked
Generative AI
Terms related to AI systems that create text, images, code, audio, video, and other content.
- AI Copilot
- Chain-of-Thought Prompting
- Code Generation
- Completion
- Content Generation
- Context Window
- Diffusion Model
- Few-Shot Prompting
- Fine-Tuning
- Foundation Model
- Generative AI
- Generative Model
- Guardrails
- Hallucination
- Image Generation
- Inference
- Large Language Model
- Multimodal Generation
- Prompt
- Prompt Template
- Retrieval-Augmented Generation
- Role Prompting
- Summarisation
- Synthetic Data
- System Prompt
- Temperature
- Text Generation
- Token
- Tokenizer
- Translation
- Zero-Shot Prompting
AI Copilot
📘 Definition
An AI assistant that supports users while they perform tasks rather than replacing them.
🪄 Simple Explanation
A copilot works alongside people.
🔑 Why It Matters
It increases productivity by offering suggestions.
💡 Example
An AI writing copilot proposes edits while you write.
See Also
Chain-of-Thought Prompting
📘 Definition
A prompting technique that encourages a model to reason through intermediate steps before producing a final answer.
🪄 Simple Explanation
It encourages the AI to work through a problem step by step.
🔑 Why It Matters
It can improve reasoning on complex tasks.
💡 Example
A prompt asks the AI to explain its reasoning before giving the answer.
Code Generation
📘 Definition
The automatic creation of computer code by a generative AI model.
🪄 Simple Explanation
AI can write software from natural-language instructions.
🔑 Why It Matters
It improves developer productivity and learning.
💡 Example
A programmer asks AI to generate a Python function.
See Also
Completion
📘 Definition
A completion is the text, code, or other content generated by an AI model in response to a prompt.
🪄 Simple Explanation
A completion is the AI's answer.
🔑 Why It Matters
Understanding completions helps users evaluate AI output.
💡 Example
After receiving a writing prompt, the AI produces a paragraph as its completion.
See Also
Content Generation
📘 Definition
The creation of text, images, audio, video, or other media using AI.
🪄 Simple Explanation
AI can produce many different kinds of content.
🔑 Why It Matters
Content generation supports education, marketing, and design.
💡 Example
An AI creates lesson plans, quizzes, and presentation slides.
See Also
Context Window
📘 Definition
A context window is the amount of information a language model can consider at one time while generating a response.
🪄 Simple Explanation
The context window is the AI's working memory for a conversation.
🔑 Why It Matters
Larger context windows allow models to process longer documents and conversations.
💡 Example
A model with a large context window can analyse an entire research paper in one request.
See Also
Diffusion Model
📘 Definition
A generative AI model that creates images by gradually transforming random noise into meaningful visuals.
🪄 Simple Explanation
Diffusion models are widely used for AI image generation.
🔑 Why It Matters
They power many modern text-to-image systems.
💡 Example
An AI artwork tool uses a diffusion model to create illustrations.
Few-Shot Prompting
📘 Definition
Providing a model with a few examples before asking it to perform a similar task.
🪄 Simple Explanation
Few-shot prompting teaches by example.
🔑 Why It Matters
It often improves consistency and accuracy.
💡 Example
Showing two correctly formatted emails before asking the AI to draft a third.
See Also
Fine-Tuning
📘 Definition
Fine-tuning is the process of adapting a pre-trained model to a specific task or domain using additional training data.
🪄 Simple Explanation
Fine-tuning specialises an existing AI model.
🔑 Why It Matters
It allows organisations to build domain-specific AI systems.
💡 Example
A hospital fine-tunes a language model using medical documents.
See Also
- Training
- Large Language Model
- Transfer Learning
Foundation Model
📘 Definition
A foundation model is a large AI model trained on broad datasets that can be adapted for many downstream tasks.
🪄 Simple Explanation
Foundation models provide a starting point for many AI applications.
🔑 Why It Matters
Most modern generative AI systems are built on foundation models.
💡 Example
A single foundation model may support translation, summarisation, and question answering.
Generative AI
📘 Definition
Generative AI refers to artificial intelligence systems that can create new content, such as text, images, code, music, video, or designs, based on patterns learned from training data.
🪄 Simple Explanation
Generative AI creates new content instead of only analyzing existing content.
🔑 Why It Matters
Generative AI is rapidly changing how people write, create images, analyse data, develop software, and solve problems. Understanding its capabilities and limitations is becoming an essential digital skill.
💡 Example
An AI tool that writes a lesson plan, generates an image, creates quiz questions, or drafts an email is using generative AI.
Generative Model
📘 Definition
A generative model learns patterns in data so it can create new content such as text, images, audio, code, or video.
🪄 Simple Explanation
A generative model creates new content rather than simply analysing existing information.
🔑 Why It Matters
Generative models are the engine behind many modern AI tools.
💡 Example
An AI image creator is powered by a generative model.
Guardrails
📘 Definition
Rules and safety mechanisms that guide AI behaviour and reduce harmful or inappropriate outputs.
🪄 Simple Explanation
Guardrails keep AI responses safer and more reliable.
🔑 Why It Matters
They are important for responsible AI deployment.
💡 Example
An AI refuses requests that violate safety policies.
See Also
Hallucination
📘 Definition
A hallucination occurs when a generative AI model produces information that is incorrect, fabricated, or unsupported while presenting it confidently.
🪄 Simple Explanation
A hallucination is an AI mistake that sounds believable.
🔑 Why It Matters
Recognising hallucinations is essential for responsible AI use.
💡 Example
An AI invents a journal article that does not exist.
See Also
- Fact Checking
- Large Language Model
- AI Ethics
Image Generation
📘 Definition
The creation of new images from text prompts or other inputs using generative AI models.
🪄 Simple Explanation
AI can create original images from descriptions.
🔑 Why It Matters
Image generation supports design, education, and creativity.
💡 Example
A teacher creates classroom illustrations by describing the scene.
See Also
Inference
📘 Definition
Inference is the process of using a trained generative AI model to produce an output from a prompt.
🪄 Simple Explanation
Inference is when a trained AI generates an answer.
🔑 Why It Matters
Most users interact only with inference, not training.
💡 Example
Typing a prompt into an AI assistant triggers inference.
See Also
Large Language Model
📘 Definition
A Large Language Model (LLM) is a deep learning model trained on vast amounts of text to understand and generate human language.
🪄 Simple Explanation
An LLM can read, write, summarise, explain, and answer questions.
🔑 Why It Matters
LLMs have transformed education, business, and software development.
💡 Example
ChatGPT is an example of an application built on an LLM.
See Also
Multimodal Generation
📘 Definition
Multimodal generation refers to creating outputs across multiple content types such as text, images, audio, or video.
🪄 Simple Explanation
Generative AI can now create more than just text.
🔑 Why It Matters
Multimodal generation expands creative and professional applications.
💡 Example
An AI generates a presentation containing text, charts, and illustrations from one prompt.
See Also
Prompt
📘 Definition
A prompt is the instruction or input provided to an AI system that guides the response it generates.
🪄 Simple Explanation
A prompt tells AI what you want it to do.
🔑 Why It Matters
Better prompts usually produce better results.
💡 Example
'Summarise this article in five bullet points' is a prompt.
Prompt Template
📘 Definition
A prompt template is a reusable prompt structure containing placeholders that can be filled with different information.
🪄 Simple Explanation
Prompt templates provide consistency.
🔑 Why It Matters
They are widely used in AI applications and automation.
💡 Example
A customer-support system inserts customer details into the same prompt template.
See Also
Retrieval-Augmented Generation
📘 Definition
A technique that combines information retrieval with generative AI so responses can use external knowledge sources.
🪄 Simple Explanation
RAG lets AI answer using trusted documents.
🔑 Why It Matters
It improves factual accuracy.
💡 Example
A company chatbot answers questions using its internal manuals.
Role Prompting
📘 Definition
A prompting technique in which the AI is assigned a specific role or persona before completing a task.
🪄 Simple Explanation
Role prompting gives the AI a clear perspective.
🔑 Why It Matters
It often produces more focused responses.
💡 Example
'Act as a financial analyst and explain this report.'
See Also
Summarisation
📘 Definition
The process of producing a shorter version of a document while preserving its key ideas.
🪄 Simple Explanation
AI can condense long documents into concise summaries.
🔑 Why It Matters
Summarisation saves time and improves understanding.
💡 Example
A research paper is reduced to a one-page executive summary.
See Also
Synthetic Data
📘 Definition
Artificially generated data that resembles real-world data and can be used for training or testing AI models.
🪄 Simple Explanation
Synthetic data is created instead of collected.
🔑 Why It Matters
It helps when real data is limited or sensitive.
💡 Example
Researchers generate synthetic medical images for model training.
See Also
System Prompt
📘 Definition
Instructions that define the overall behaviour, style, or constraints of an AI assistant before user prompts are processed.
🪄 Simple Explanation
A system prompt sets the AI's overall behaviour.
🔑 Why It Matters
It helps maintain consistency and safety.
💡 Example
A chatbot is instructed to respond using formal academic language.
See Also
Temperature
📘 Definition
A model parameter that controls the randomness of generated output.
🪄 Simple Explanation
Temperature adjusts how creative or predictable AI responses are.
🔑 Why It Matters
Choosing the right temperature improves response quality.
💡 Example
A low temperature gives consistent answers, while a high temperature encourages creativity.
See Also
Text Generation
📘 Definition
The production of written content such as articles, emails, reports, or summaries by a language model.
🪄 Simple Explanation
AI can generate many forms of written content.
🔑 Why It Matters
Text generation is one of the most common AI applications.
💡 Example
An AI drafts a business proposal from bullet points.
See Also
Token
📘 Definition
The smallest unit of text processed by a language model. Words may consist of one or more tokens.
🪄 Simple Explanation
Tokens are the pieces of text AI reads.
🔑 Why It Matters
Token limits determine how much information a model can process.
💡 Example
A long report is split into thousands of tokens before processing.
See Also
Tokenizer
📘 Definition
A tokenizer converts text into tokens that can be processed by a language model.
🪄 Simple Explanation
A tokenizer breaks text into pieces the AI understands.
🔑 Why It Matters
Tokenisation is the first step in language processing.
💡 Example
The sentence 'AI is useful' is split into several tokens.
See Also
Translation
📘 Definition
The conversion of text from one language to another using AI.
🪄 Simple Explanation
AI can translate between many languages.
🔑 Why It Matters
Translation supports global communication and learning.
💡 Example
A student translates lecture notes from French into English.
Zero-Shot Prompting
📘 Definition
Zero-shot prompting asks an AI to perform a task without providing examples.
🪄 Simple Explanation
The AI completes a task using only your instructions.
🔑 Why It Matters
It demonstrates the knowledge already contained in the model.
💡 Example
'Summarise this article' is a zero-shot prompt.