Narrow AI

What is Narrow AI?

Narrow AI is a term used to describe artificial intelligence systems that are specified to handle a singular or limited task. The antithesis to Narrow AI, sometimes referred to as weak AI, is called strong AI. Strong AI, unlike Narrow AI, is capable of handling a wide range of tasks rather than one particular task or problem. This variation of artificial intelligence can be roughly conceptualized as a foundation for neural networks emulating sentience or consciousness.


Applications of Narrow AI

Applications of Narrow AI are becoming increasingly common as machine learning is continually integrated into modern society. For example, Narrow AI may be used for spam email filtering, music recommendation services, and even autonomous vehicles. There are some concerns, however, about the widespread use of Narrow AI in important infrastructure functions. Some argue that characteristics of Narrow AI make it brittle, and in cases in which a neural network may be used to control important systems (e.g. power grid, financial trading) alternatives may be more risk averse.

Examples of Narrow AI

One of the most availably integrated forms of Narrow AI is Apple's Siri. Siri is a narrow artificial intelligence algorithm that brings the functions of machine learning to the mobile platform of an iPhone. While Siri is helpful at completing various specific tasks, it is by no means a strong AI, and often has challenges with tasks outside its range of abilities. Because Siri expresses no self-awareness or genuine intelligence, it is a foundational example of Narrow AI.

A popular contestant on the television show Jeopardy!, IBM's Watson is technically a complex incarnation of Narrow AI. Like Siri, Watson is not truly sentient, but does integrate Narrow AI well enough to cover a range of functions.