- Define artificial intelligence and distinguish it from related concepts such as machine learning and deep learning
- Trace the history of AI from the 1956 Dartmouth workshop through two AI winters to the modern deep learning era
- Distinguish between narrow AI (ANI), artificial general intelligence (AGI), and artificial superintelligence (ASI)
- Identify the three classical paradigms of machine learning (supervised, unsupervised, and reinforcement learning) and the rise of self-supervised learning
- Describe the stages of the end-to-end AI pipeline, from problem definition through deployment and monitoring
A clinical resident in Auckland uploads a chest radiograph and a model returns a probability of pneumonia within two seconds. A teenager in Lagos uses a chatbot to draft a university entrance essay and to negotiate the structure of an argument. A protein engineer in Cambridge submits an amino-acid sequence and receives a three-dimensional structure that, a decade ago, would have demanded a year of crystallography. A driver on the 405 freeway in Los Angeles takes their hands off the wheel as the car merges, brakes, and changes lanes on its own. None of these things would have been possible in 2010. By early 2026 they are unremarkable.
Artificial intelligence has moved, in less than a generation, from the philosophy seminar to the operating theatre, the design studio, the search engine, the warehouse, and the classroom. It sits inside applications that several billion people open daily. It is also sitting at the centre of one of the largest capital reallocations in technological history: hundreds of billions of dollars in data-centre construction, semiconductor fabrication, and electricity supply, all driven by the belief that machine intelligence is now becoming a general-purpose technology of the same order as electricity or the internal combustion engine.
This chapter has two aims. The first is a careful working definition of what AI is and is not. The second is to place the modern field in the context of its own history, so that the techniques you will study later (search, logic, probabilistic graphical models, neural networks, reinforcement learning, transformers, large language models) appear not as isolated tricks but as the latest layer of a seventy-year intellectual project. The technical machinery comes in later chapters; here we are mapping the territory.