Narrow AI is artificial intelligence designed to perform a specific task, or a tightly related group of tasks. It is also called weak AI or artificial narrow intelligence (ANI). It can be extremely good at its job, but it cannot move that skill to an unrelated problem on its own.
What makes it “narrow”
“Narrow” refers to scope, not to power. A chess engine can beat every human grandmaster, yet it cannot book a flight or read an X-ray. It was built, trained and tested for one domain. Outside that domain, it has no ability at all.
Examples of narrow AI
- Spam filters that sort your inbox
- Face unlock on your phone
- Product recommendations on shopping sites
- Navigation apps predicting traffic
- Voice assistants that set timers and answer simple questions
- IBM’s Deep Blue, which beat chess champion Garry Kasparov in 1997
Are chatbots narrow AI?
Large language models blur the line. One model can write code, summarise a report and translate a paragraph. That is far broader than a spam filter. Still, most researchers classify them as narrow AI because they lack reliable reasoning across all domains, do not learn continuously from experience and can fail badly on tasks outside their training patterns. Some researchers argue they are an early, partial step toward general AI. The debate is open.
Narrow AI vs general AI
Narrow AI solves defined problems. General AI would learn and adapt to almost any intellectual task, the way a person can move from cooking to accounting. Narrow AI is here and in daily use. General AI remains a research goal.
Why it matters for business
Because narrow AI is task-specific, it works best when the task is clearly defined, the data is good and the cost of an error is known. Most successful AI projects are narrow on purpose.