Definitions and Terminology
Artificial Intelligence (AI)
Software that can produce outputs (such as predictions, recommendations, or decisions) based on data and objectives set by humans. AI systems typically infer patterns from inputs and use those patterns to generate an output.
Model
A component of an AI system that transforms inputs into outputs (for example, producing text, classifications, or recommendations). Models are usually created using statistical or machine-learning techniques.
Generative AI (GAI)
A class of AI models designed to generate new content (such as text, images, audio, or video) based on patterns learned from data.
GAI includes text-based models used by chatbots (for example ChatGPT, Gemini, and Claude) and may also be embedded within other products we use (for example Google Workspace or project-management tools).
GAI also includes image, voice, and video generation tools (for example DALL-E, Adobe Firefly, or text-to-speech tools).
AI tool
An end-user product or feature that uses AI or GAI (for example a chatbot app, a coding assistant, or an “AI” feature inside another product).
Provider
The organisation that offers an AI model or AI tool (for example OpenAI, Anthropic, Google, Microsoft, Mistral).
Prompt
The instructions and any context a user provides to an AI tool. Prompts may include text, images, files, or structured data.
Model Stability (output consistency)
How consistent and reliable an AI tool’s outputs are when prompts, context, or conditions change. More mature tools often behave more predictably due to wider testing and operational controls, but any output may still require verification.
Last updated: