Dar Yapay Zeka
Specifically, specialized in a limited task, the type of artificial intelligence that can not be generalized outside of that area.
Dar artificial intelligence (Narrow AI or Artificial Narrow Intelligence, ANI), defines the type of artificial intelligence, which shows high qualification in a certain, limited task or narrow area, but when out of this field, it can not be generalized. Artificial intelligence that plays a chess can be on top level but using this information can not give a recipe proposal or does not take a car while; a face recognition system can recognize the faces extremely hit but cannot translate language. This is the basic feature that distinguishes narrow artificial intelligence from General Artificial Intelligence (Artificial General Intelligence, AGI), a concept that has the flexibility and versatility of human intelligence in theory.
Today, almost all artificial intelligence systems used for commercial and research purposes, see how impressive they appear, technically enters the category of narrow artificial intelligence: image recognition systems, recommendation engines, chat bots, or even large language models can not be able to reliably exclude data distribution and task-specific molds. The concept of narrow artificial intelligence is important to distinguish the actual progress in the field of artificial intelligence from exaggerated expectations; a model does not mean that in a particular task, it does not mean that the model has a general understanding of human-like or known, only it indicates that it is a statistical image recognition system optimized for narrow task.
Today, almost all artificial intelligence systems used for commercial and research purposes, see how impressive they appear, technically enters the category of narrow artificial intelligence: image recognition systems, recommendation engines, chat bots, or even large language models can not be able to reliably exclude data distribution and task-specific molds. The concept of narrow artificial intelligence is important to distinguish the actual progress in the field of artificial intelligence from exaggerated expectations; a model does not mean that in a particular task, it does not mean that the model has a general understanding of human-like or known, only it indicates that it is a statistical image recognition system optimized for narrow task.