You are not only confusing cognitive AI with symbolic AI, you are also forgetting the huge differences between today's Deep Learning and the marvels of the human brain - we learn incrementally from much smaller datasets and are good at generalisation. A case in point is how small children rapidly learn language from their parents and peers.

Cognitive AI seeks to mimic human thought (memory, reasoning and learning) inspired by advances in the cognitive sciences and over 500 million years of neural evolution. This involves a combination of symbolic and statistical approaches that is complementary to neural networks. For more details see the W3C Cognitive AI Community Group.