Weak AI, also known as narrow AI, refers to artificial intelligence systems that are designed to handle specific, limited tasks. Unlike strong AI, which aims to perform a wide range of cognitive tasks at human levels or beyond, weak AI focuses on particular areas where it can perform as well or better than humans. Examples of weak AI include voice recognition systems like Siri and Alexa, chess-playing programs, and even more sophisticated systems like self-driving cars.

Weak AI operates under a constrained set of guidelines and doesn't possess consciousness or self-awareness. It's essentially programmed to be very good at specific tasks, operating within predefined rules or learned patterns. This type of AI has made significant impacts across various sectors, helping to optimize and automate processes but is limited by its inability to transcend its programming for generalized understanding or reasoning.

The term "weak AI" was popularized by philosopher John Searle, who contrasted it with strong AI—the concept of a machine with the ability to understand and reason on a par with a human being. Searle's distinction emphasizes that while weak AI can simulate human cognition and perform tasks requiring human-like intelligence, it does not have a mind of its own in the way strong AI theoretically would.