Essential artificial intelligence terms explained clearly, in English. For professionals, managers and teams who want to understand the technology they already use every day.
A field of computer science dedicated to creating systems capable of performing tasks that normally require human intelligence, such as language understanding, pattern recognition and decision-making.
An autonomous program that executes tasks, makes decisions and interacts with tools on behalf of a user, based on defined objectives.
The ability to understand how AI systems work, their risks and limitations, and how to use them responsibly. Required by Article 4 of the EU AI Act.
An interface that enables communication between different software systems. It defines rules and formats so that applications can exchange data in a structured way.
The execution of tasks or processes without direct human intervention, through rules or artificial intelligence.
A program that simulates conversations with users through text or voice, using predefined rules or language models.
A computing model that provides access to resources (servers, storage, applications) via the internet, without local infrastructure.
A subfield of machine learning that uses neural networks with multiple layers to learn complex patterns in large volumes of data.
Regulation (EU) 2024/1689 of the European Parliament, the first comprehensive legal framework for artificial intelligence. It partially entered into force on 2 February 2025.
The process of adapting a pre-trained AI model to a specific task or domain, using additional, more focused data.
A language model architecture developed by OpenAI, trained to generate text. The foundation of ChatGPT.
An AI system trained to understand and generate text in natural language.
A large-scale language model, trained on large volumes of text, capable of generating, summarising and understanding natural language. Examples: GPT-4, Claude, Gemini, LLaMA.
A subfield of artificial intelligence in which systems learn from data, identifying patterns and improving performance without being explicitly programmed for each task.
An area of AI dedicated to the interaction between computers and human language: understanding, generation and translation of text.
Software whose source code is public and can be freely used, modified and distributed, within the terms of the licence.
An instruction or input text given to an AI model to obtain a specific response or result.
A technique that combines information retrieval from documents with text generation by an LLM, enabling responses based on specific, up-to-date data.
Classification of AI systems into four categories: unacceptable risk (prohibited), high, limited and minimal. It defines the applicable regulatory obligations.
A software distribution model where the application is accessed via the internet, with subscription-based payment, without local installation.
The basic unit of text processed by a language model. It can be a word, part of a word or a punctuation character. Models have token limits per request.
A specific sector or industry (e.g. healthcare, real estate, fitness). At D'One, agents are configured per vertical to understand the particularities of each sector.
D'One sets up autonomous AI agents for SMEs — scheduling, invoicing, content and more. Fully managed, fully compliant with the EU AI Act.
Get in touch