Narrow AI refers to systems engineered to perform a specific task or bounded set of tasks, achieving strong performance within their domain without the general adaptability associated with artificial general intelligence.
Semantic Classification
Content
- Narrow AI covers nearly all deployed systems, including image classifiers, recommendation engines and speech recognisers, each optimised for a particular objective. Performance does not transfer beyond the domain the system was trained for.
- The term contrasts with artificial general intelligence, which would generalise across tasks at human breadth. Most metaverse and XR applications use narrow models for vision, tracking and content generation rather than general systems.