AI glossary

Core terms of artificial intelligence — explained clearly, kept current and linked to the reporting.

• AI agentAn AI agent uses a language model to pursue goals autonomously across multiple steps – with planning, tool calls and feedback loops.
• Context windowThe context window is the maximum amount of text in tokens a language model can process per call – it bounds knowledge, history and cost.
• EmbeddingEmbeddings are vector representations of content that map semantic similarity to spatial proximity – the basis of semantic search and RAG.
• EU AI ActThe EU AI Act regulates AI systems on a risk basis – with staggered obligations from banned practices to high-risk requirements from 2026/27.
• Fine-tuningFine-tuning adapts a pre-trained language model with task-specific data – suited to style, format and domain language, less to factual knowledge.
• Generative Engine Optimization (GEO)Generative engine optimization (GEO) prepares content to be cited in answers from AI search engines and chatbots – SEO for the AI era.
• HallucinationHallucinations are plausible-sounding but false outputs of language models – the central risk in production use of generative AI.
• Large Language Model (LLM)Large language models are neural networks that model language statistically and generate text – the foundation of modern generative AI systems.
• Model Context Protocol (MCP)The Model Context Protocol (MCP) is an open standard that connects AI applications to external tools and data sources in a uniform way.
• Retrieval-Augmented Generation (RAG)Retrieval-augmented generation enriches language model answers with externally retrieved knowledge – the standard for current and internal data.