Fundamentals
The foundation vocabulary to discuss AI in the executive committee. Without these shared terms, the rest of the glossary only reads halfway.
- AGI (artificial general intelligence)AGI (artificial general intelligence) refers to an AI system able to match or exceed humans at most cognitive tasks, not just one. There is neither a shared definition nor an accepted test to say it has been reached: each lab uses its own.
- Artificial intelligence (AI)Artificial intelligence (AI) refers to the set of techniques that enable a machine to perform tasks associated with human intelligence: recognising an image, understanding a text, predicting, deciding, writing. Under EU law, an “AI system” is identified by its capacity to infer: it infers from the input it receives how to generate predictions, content, recommendations or decisions (AI Act, Article 3).
- Computer visionComputer vision is the field of artificial intelligence that enables a machine to analyse images and video: recognising objects, detecting defects, reading a document, tracking movement. In Europe, its biometric uses (facial recognition, emotion recognition) are partly prohibited and partly classified as high-risk by the AI Act.
- Deep learningDeep learning is a form of machine learning that uses artificial neural networks made up of many layers, each processing data at a higher level of abstraction. It drives the progress in image recognition, speech recognition, translation and generative AI.
- EmbeddingAn embedding is the numerical representation of a word, a sentence, or a document, in the form of a vector in a space of several hundred or thousand dimensions. Two texts semantically close have geometrically close embeddings, which enables search by meaning rather than by keywords.
- Foundation modelA foundation model is a generalist AI model trained on massive, diversified data, which serves as a reusable base for dozens or even hundreds of different business applications, by adaptation (fine-tuning, RAG, prompt) rather than by specific training.
- Generative AIGenerative AI is a family of artificial intelligence models capable of producing new content (text, image, sound, video, code) from a natural-language instruction, as opposed to traditional AI which was limited to classifying, predicting, or detecting.
- InferenceInference is the usage phase of an AI model, during which the model computes a response from a given prompt. It is the operation billed by API providers, distinct from training which is an initial fixed cost.
- LLM (Large Language Model)An LLM (Large Language Model) is a type of artificial intelligence trained on text corpora of several hundred billion words, which produces natural language by predicting, word by word, the most probable continuation of a given text.
- Machine learningMachine learning is the branch of artificial intelligence in which a program learns to perform a task from examples, instead of following rules written by a programmer. It relies on a model whose parameters are automatically adjusted on data until it produces reliable results.
- MultimodalA multimodal AI model is a model capable of processing and producing several types of content simultaneously: text, image, audio, video, code. The same model can analyse a photo, understand a voice query, read a document, and reply in writing, without any intermediate pipeline.
- Neural network (artificial)An artificial neural network is a computing model made of small interconnected units, artificial neurons, organised in layers: each neuron combines its inputs according to weights, then passes a signal to the next layer. Learning consists of automatically adjusting these weights, from examples, until the network produces the right answers.
- NLP (natural language processing)NLP (natural language processing) is the field of artificial intelligence that enables a machine to analyse, understand and produce human language, written or spoken. Large language models (LLMs) such as ChatGPT or Claude are now its most visible form.
- OverfittingOverfitting describes an AI model that fits its training data so closely that it memorises it instead of learning general rules: it scores very well in internal tests, then makes mistakes on new cases. It is one of the most common reasons why a model that shines in a demo disappoints in production.
- ParameterA parameter of an AI model is a numerical value learned during training that determines the strength of connections between the artificial neurons of the network. The parameter count measures the model's capacity to store associations between the elements of its training corpus.
- Reasoning modelA reasoning model is a large language model trained to produce internal, step-by-step thinking before giving its answer. It is more reliable on complex problems (analysis, calculation, code, agent planning), but slower and more expensive, because its thinking tokens are billed.
- Reinforcement learningReinforcement learning is a machine learning method in which a system learns which actions to choose by trying them and receiving a reward or a penalty, so as to maximise a cumulative gain over time. Long associated with games and robotics, it is also used to adjust the behaviour of conversational assistants (RLHF) and to train reasoning models.
- RLHF (reinforcement learning from human feedback)RLHF (reinforcement learning from human feedback) is a training technique that adjusts a language model according to the preferences of people who compare and rank its answers. Popularised by InstructGPT (OpenAI, 2022) and then ChatGPT, it turned models that continued text into assistants that follow instructions, with a documented side effect: a tendency to agree with the user.
- Supervised learningSupervised learning is a machine learning method in which a model is trained on examples whose correct answer is known (a label), so that it can predict that answer for new cases. It is found in common predictive uses in business: customer scoring, fraud detection, automatic sorting of requests.
- TokenA token is the elementary unit of text that an AI model manipulates, generally a word fragment of 2 to 4 characters, depending on the model and the language. It is both the model's unit of computation and the AI providers' unit of billing, counted separately on input and output.
- TrainingTraining is the construction phase of an AI model, during which the model ingests a massive corpus and adjusts its billions of internal parameters to learn the statistical regularities of that corpus. It is the most expensive operation in a model's life cycle.
- TransformerThe transformer is the neural-network architecture, introduced by Google in 2017, that underpins nearly all current generative AI models. Its central innovation is the attention mechanism, which lets the model dynamically weigh the relative importance of words in a sequence.
- Unsupervised learningUnsupervised learning is a machine learning method that looks for structure in raw data, without any expected answer provided in advance: groups of similar items, unusual behaviour, variables that go together. It is mainly used to explore and segment (customers, products, transactions), and its results must be interpreted by people who know the business.