Foundations
Artificial IntelligenceAI
Also known as: machine intelligence
The field of building computer systems that perform tasks we normally associate with human intelligence, such as understanding language, recognising images or making decisions.
Artificial intelligence, or AI, is the effort to make computers do things that normally need human thinking, like understanding what you say, recognising a face in a photo, or suggesting the fastest route home. It is not one single invention but a whole field of technology. And it is not a robot with feelings: today's AI has no awareness or intentions, it is very advanced pattern-matching that can be remarkably useful.
Artificial intelligence is the branch of computer science concerned with building systems that carry out tasks that would normally require human intelligence: perceiving, reasoning, learning, understanding language and making decisions. The term dates back to a 1956 workshop at Dartmouth College, where the field got its name.
A field, not a single thing
"AI" covers many approaches. Early systems were built on hand-written rules. Most of today's progress comes from machine learning, where systems learn patterns from data rather than being explicitly programmed. The recent wave of chatbots and image generators is a subset called generative AI.
Narrow vs general
Almost all AI in use today is narrow: very capable at specific tasks but not generally intelligent. A hypothetical system that matches humans across the full range of thinking is called artificial general intelligence (AGI), which does not exist today and remains a topic of research and debate.
Artificial intelligence is the study and engineering of agents that perceive an environment and act to achieve objectives, formalised in modern texts as the design of rational agents (Russell & Norvig). Historically the field oscillated between symbolic approaches (logic, search, expert systems, knowledge representation) and statistical/connectionist ones (neural networks). The current paradigm is dominated by data-driven machine learning, and in particular deep learning: multi-layer neural networks trained by gradient descent on large datasets.
The distinction between narrow AI (task-specific, the entirety of deployed systems) and artificial general intelligence remains central, as do the framing questions of capability, reliability and alignment. It is worth separating the scientific field from the marketing label: "AI" today most often refers to specific statistical models, most prominently large language models, rather than any system possessing understanding or agency in the human sense.