What is Artificial General Intelligence?
Artificial General Intelligence (AGI) refers to a theoretical future AI system that possesses the ability to understand, learn, and apply knowledge across a wide variety of independent tasks at a level equal to or beyond human capabilities.
How does it differ from current AI?
Every AI system deployed today is "narrow" AI. A narrow AI can be superhuman at a specific task—like playing chess, translating text, or identifying diseases in X-rays—but it is completely useless at anything else. A chess AI cannot tell a joke or drive a car. AGI describes a single system that can learn to do all of those things autonomously, adapting to novel situations it was never explicitly trained for.
What is a simple example?
If you hire a plumber, they know how to fix pipes (narrow AI). If you hire a brilliant polymath who can read a manual and immediately learn how to fix pipes, perform surgery, write a symphony, and negotiate a treaty, that represents the flexibility of AGI.
Why does it matter?
AGI is the ultimate, stated goal of many leading artificial intelligence research labs. The realization of AGI would represent a profound technological paradigm shift, as it could theoretically automate almost all cognitive labor and dramatically accelerate scientific research.