Artificial intelligence
Introduit| English | Sens |
|---|---|
| AI/ˌeɪ ˈaɪ/ | IA |
| narrow AI/ˈnærəʊ ˌeɪ ˈaɪ/ | IA étroite |
| general AI/ˈdʒenərəl ˌeɪ ˈaɪ/ | IA générale |
| strong AI/strɒŋ ˌeɪ ˈaɪ/ | IA forte |
| expert systems/ˈekspɜːt ˈsɪstəmz/ | systèmes experts |
| natural language processing/ˈnætʃərəl ˈlæŋɡwɪdʒ ˈprəʊsesɪŋ/ | traitement du langage naturel |
Machines that seem to think
- AI 人工智能 is the simulation of human intelligence by computer systems.
- It means a computer doing tasks that normally need human thinking.
- AI is grouped into three levels — only one of which exists today.
Laboratoire de concepts informatiques
Classify concrete examples by the computing idea they demonstrate.
Characteristics of AI
- AI systems collect data and the rules for using it.
- They can reason (work things out with the rules) and draw conclusions — which may be approximate (a best guess) or definite.
- They can learn from data and adapt — changing their behaviour as they get new data.

Artificial intelligence is:
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
AI is a computer doing tasks that normally need human thinking.
Which is a characteristic of an AI system?
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
AI collects data + rules, reasons, draws conclusions, learns and adapts.
Examples and levels of AI
- Examples: expert systems 专家系统 (advice like a human expert, e.g. medical diagnosis), natural language processing 自然语言处理, and self-driving cars.
- Narrow AI — does only one task or a small set (e.g. a chess program). All AI today is narrow AI.
- General AI — could do any task a human can. Strong AI — would think and be self-aware like a human mind. Neither exists yet.


Match each AI term to what it means.
Difficulté 2/5 · 140 pièces + un petit bonus lors de la première réussite aujourd'hui
AI simulates thinking; expert systems advise from rules; today's AI is all narrow; strong AI does not yet exist.
Narrow AI:
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
Narrow AI is task-specific (e.g. chess); general and strong AI do not exist yet.
All AI in use today is narrow AI (good at one specific task); strong AI, with human-like self-awareness, does not yet exist.
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
A chess engine or a recommender is narrow AI; a self-aware, general mind (strong AI) remains theoretical.
The limits of AI
AI is only as good as its data. A system trained on biased or incomplete data makes biased decisions. AI can recognise patterns superbly but does not truly "understand" — judgement and oversight still matter.

Worked example: an expert system for car faults
A garage uses an expert system to find car faults. Describe how it reaches a diagnosis.
- The mechanic answers questions through the user interface, for example "Does the engine turn over? No." These answers are the facts. The knowledge base holds facts from human experts, and the rule base holds IF–THEN rules, such as IF the engine does not turn over AND the lights are dim THEN the battery is flat.
- The inference engine applies the rules to the facts and gives a diagnosis, often with a probability, such as "battery flat: 85%". It is narrow AI: it is good at one job only.
Which part of an expert system applies the rules to the facts?
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
The inference engine reasons with the rule base to reach a conclusion.
A self-driving car is an example of general AI.
Difficulté 1/5 · 100 pièces + un petit bonus lors de la première réussite aujourd'hui
It does one set of tasks (driving), so it is narrow AI; general AI does not exist yet.
You've got it
- AI = a computer simulating human intelligence
- AI characteristics: collect data + rules, reason, draw conclusions, learn, adapt
- examples: expert systems, NLP, self-driving cars
- narrow (all AI today) → general → strong (self-aware, not yet real)