Synonyms:
Weak artificial intelligence, Artificial narrow intelligence

What is Narrow AI?

Narrow artificial intelligence (ANI), also known as weak AI, has very limited capabilities, compared with what is to come. It is designed for specific purposes, and performs practical tasks.

The narrow version of AI is inferior to human intelligence. It can do its one or a few demarcated tasks – usually much better, faster or more accurate than a human, but it is in principle still restricted to these tasks.

Narrow AI has limited capabilities

Still a one-trick pony with limited leeway.

These weak AI models are not sentient (self-aware). They lack both consciousness, and the ability to reason independently. The next phase is General AI, which would have these abilities.

See also: Artificial intelligence in a nutshell and General artificial intelligence.

WHERE ARE WE NOW?

At present, all known artificial intelligence is narrow, although many AI models are gaining complexity.

EXAMPLES NARROW AI

Examples of narrow AI are self-driving cars, facial recognition tools, or virtual assistants like Alexa, Google Assistant and Siri. Impressive as these may be, they are still forms of weak AI.

See also: Unsupervised learning and Superintelligence: a promise or a threat?

FOOD FOR THOUGHT

It is good to realize that the impressive AI algorithms that are on the market right now are still very basic narrow AI. AI technology is still in its infancy.

It is also important to realize that AI is an crucial strategic asset in the geopolitical arena, so many projects are shrouded in secrecy.