AI is all about searching things like finding patterns or finding ways of clustering things. Most of the AI performs those searches based on the data they are fed with. Without past data an AI cannot take any decision, so there is always some degree of learning involved.
An exception are the old fashioned AI systems which are based on rules. Those AIs do not need (mostly) data to learn how to take decisions. They need, however, a domain expert which produces the knowledge used by the AI to take decisions. For example, the AI needs to know that rain makes things wet to decide that it should not stay under the rain in order not to be wet.