Skip to content

91 concepts to understand artificial intelligence

Eight chapters, eight entries to read

Artificial intelligence answers our questions, completes our sentences, creates images and organizes our access to information. It is everywhere, but what do we really know about how it works? What is a model? How does a machine learn? What are a token, a vector, fine-tuning or a hallucination?

This living lexicon offers a clear, rigorous and accessible introduction, free of fantasies and easy fears. Organized in thematic chapters, it reads like a story: from the most fundamental to the most applied, from inside the models to their effects on the world.

The eight chapters

1. What Artificial Intelligence Is

Before getting into technical detail, let’s start with the basics. What is artificial intelligence? (15 concepts)

Imitation Game · Artificial Intelligence · Generative Artificial Intelligence · Algorithm · Model · Phases (of a Model) · Training · Inference · Fine-Tuning · Foundation Model · Agent · Alignment · Transparency · Explainability · Open Source

2. What an AI System Contains: Neurons and Parameters

Inside every artificial intelligence lies a complex mathematical structure: neurons, weights, functions, layers, and so on. (16 concepts)

Perceptron · Neuron · Parameter · Weight · Bias (Parameter) · Activation Function · Layer · Neural Network · Backpropagation · Attention · Transformer · Language Model · Large Language Model · Reasoning Model · Diffusion Model · Multimodal

  • Excerpt to read: Neuron · page 40

3. What (and How) AI Learns: Forms of Learning

An AI system is not programmed in the classic sense of the word: it learns. But what does learning mean for a machine? (11 concepts)

Learning · Machine Learning · Supervised Learning · Unsupervised Learning · Reinforcement Learning · Deep Learning · Dataset · Training Data · Personal Data · Pattern · Overfitting

4. What AI “Understands”: Language, Images, Sound

Artificial intelligence does more than process numbers: it handles language, analyzes images, recognizes voices. (9 concepts)

Natural Language Processing · Chatbot · Generative Pre-trained Transformer · Machine Translation · Speech Recognition · Text-to-Speech · Computer Vision · Facial Recognition · Image Generator

5. The Language of Models: Tokens and Computation

What does an AI system “see” when it reads a word, an image or a sentence? It does not understand the way a human does: it computes. (6 concepts)

Token · Context Window · Vector · Probability · SoftMax · Temperature

  • Excerpt to read: Token · page 83

6. What AI Produces: Answers, Errors, Hallucinations

AI answers, predicts, generates. But it can also get things wrong, distort or invent. (12 concepts)

Input / Output · Prompt · Chain of Thought · Retrieval-Augmented Generation · Context Drift · Hallucination · Algorithmic Bias · Over-elaboration · Sycophancy · Black Box · Prompt Injection · Deepfake

7. AI in Our Digital Lives: Filters, Engines, Interfaces

Artificial intelligence structures our everyday digital uses: it filters, ranks, recommends, moderates. (10 concepts)

Personalization · News Feed · Recommender System · Collaborative Filtering · Ranking · Content Moderation · Spam Detection · Filter Bubble · Latency · Vibe Coding

8. Where Is Artificial Intelligence Going? Horizons, Risks, Questions

Artificial intelligence is advancing quickly, and with it come decisive questions. (12 concepts)

Computing Power · Agentic AI · Artificial General Intelligence · Consciousness (Sentience) · AI Scheming · AI Psychosis · Existential Risk · AI Ecological Impact · AI Race · Sovereign AI · AI Governance · Responsible AI

How to cite this book

Vial, S. (2026). A Living Lexicon of Artificial Intelligence. Stéphane Vial, publisher, Montréal.

Book cover: on a coral background, the title “A Living Lexicon of Artificial Intelligence” in deep blue, and the name Stéphane Vial
Download the cover (JPG, 212 KB)