Cognitive science and neural networks drive AI evolution

The development of artificial intelligence is deeply rooted in cognitive science rather than just traditional computer science. While early AI concepts focused on machines following specific rules, a significant shift occurred in 1958 when psychologist Frank Rosenblatt developed the Perceptron. This model was designed to learn from examples using artificial neural networks, an approach modeled after the architecture of the human brain and inspired by psychological research into how neural connections strengthen.

Despite these biological inspirations, debate continues regarding whether Large Language Models (LLMs) possess true understanding. Critics argue that these systems may simply be translating patterns of symbols into other patterns of symbols—much like a person matching foreign characters in a book without knowing the language—rather than engaging in genuine cognitive thought or linguistic comprehension.