AI behaves more like a brain than a database – cognitive science’s role in its origin story helps explain why
Modern AI owes its existence to both cognitive and computer science. What you envision as its origin story can influence how you understand what large language models can do.
Large language models, such as ChatGPT, often spark thoughts of computers pulling up files to provide responses. However, this analogy is flawed. The true origins of AI stem from cognitive science rather than just computer science. Understanding this distinction sheds light on why these systems behave the way they do.
The term "artificial intelligence" was introduced in 1956 at a Dartmouth College workshop. Initially, researchers considered a machine intelligent if it followed a set of specific rules. However, psychologist Frank Rosenblatt took a different approach in 1958. Instead of creating a machine that followed predefined rules, he built the Perceptron—a computer algorithm modeled loosely after the human brain's neural connections.
This approach laid the groundwork for machine learning by drawing inspiration from cognitive scientists like Donald Hebb, who studied how neural connections strengthen when used.
Further advancements emerged in the 1980s, when cognitive and computer scientists collaborated to enhance artificial neural networks with multiple layers. This concept, known as "deep learning," allowed these systems to tackle more complex tasks and generalize their knowledge to new examples. As computing power increased, graphics chips, the transformer architecture, and other technologies were developed.
However, it was the foundational ideas rooted in cognitive science—learning from examples and basing architecture on the human brain—that truly propelled AI forward.
These origins matter practically because modern AI systems are "grown" from examples rather than following strict rules. This growth leads to outputs that are less predictable and more prone to confabulation—making up information on the fly. Considering AI as more akin to a primitive brain, rather than a deterministic database, helps explain its unpredictable "behavior."
For instance, AI can provide different answers to the same question, much like how human memory can vary based on context and experience. Additionally, AI's probabilistic nature, similar to human memory, means it may not always perceive the world in the same way as humans do. This understanding is crucial for interpreting AI's outputs accurately, especially when those outputs might contain inaccuracies or "hallucinations."
Written by urgent.news from The Conversation's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.
