On Artificial Intelligence
And why we need to study Human Competence
A Brief History of AI
Definition 1: Machines that solve problems the way humans solve problems
The first attempts at Artificial Intelligence (AI) in the mid 20th century, attempted to mimic the human brain. Early Computer Scientists were fascinated at how simple artificial neural networks, which had a general resemblance to the neurones and synaptic connections of the brain, could solve simple problems. Many hailed a new age, and a new type of computer which could end-up solving all problems humans could solve, and attain and surpass human intelligence.
This euphoria was short-lived and died down. Firstly, due to hardware constraints, early AI systems could not scale beyond toy problems. Secondly, there was more and more debate and controversy on how exactly human intelligence worked, and how humans solve problems. Humans had little consensus on what “Human Intelligence” meant, let alone the “Artificial” variety. Finally, there were systems which bore no resemblance to the brain what-so-ever but could still solve interesting and “intelligent” problems.