The Basics of Artificial Intelligence: A Guide for Beginners
A spam filter, a search engine, and a program that drafts a letter all use forms of artificial intelligence. They perform different tasks, but each gives us a starting point for understanding what AI can do. The first step is to distinguish a system built for a particular task from the broader idea of general intelligence.
Artificial Intelligence: An Overview
At its core, Artificial Intelligence refers to the attempt to simulate human intelligence processes by machines, like computers. These processes include the ability to learn, reason, and self-correction (1).
AI is not just about programming a computer to perform tasks. It is about creating systems that can "think" and make decisions like a human. AI is often broken down into two types: Narrow AI and General AI.
Narrow AI
Narrow AI (sometimes called Weak AI) is designed to perform a specific and narrow task, such as voice recognition, recommendation systems, or image recognition. It operates under defined constraints and is focused on a single task. For instance, Siri, the virtual assistant on Apple devices, is an example of Narrow AI. Siri can help you find information, get directions, send messages, and more, but it operates under a limited set of constraints and is focused on a narrowly defined set of tasks (2).
General AI
Artificial general intelligence, or AGI, refers to the idea of an AI system with broad capabilities across many kinds of tasks, often compared with human performance. Researchers use different definitions and measures of that breadth and ability. Assessing a claim about AGI therefore requires asking which tasks were tested, how well the system performed, and what limitations remain.
These distinctions matter when choosing a tool. Ask what task it was built to perform, what information it needs, and how you will check its work. A useful result in one setting does not tell you how it will perform in another.
