THE TURING TEST
The world of artificial intelligence really got off the ground in the early 1900’s when Leonardo Torres y Quevedo created the first ever chess automaton, and Konrad Zuse built the first working program-controlled computers. However, it wasn’t until the year 1950 that a man named Alan Turing invented the Turing Test. This test was designed to provide a practical definition of intelligence when dealing with machines. A computer would pass this test if a human, after asking the computer various questions, could not tell whether the responses were coming from a person or a computer. It was questionable during this time if that meant the program was truly intelligent.
Now in order for an AI agent to pass the Turing Test it would need to be able to communicate successfully in English, be able to store information that it took in, a basic level of reasoning in order to draw conclusions, and an ability to adapt to new circumstances and pattern recognition. The Turing Test did avoid direct physical contact between the human and the computer because it was said that physical simulation was unnecessary for true intelligence but to truly pass this test the computer would need vision and robotics to manipulate objects.
Nevertheless, despite the fact that the Turing Test remains relevant today researchers haven’t been all that focused on passing it because a lot of them don’t like the idea of simply striving to emulate humans. After all, the Wright brothers didn’t succeed in flying by imitating birds and many current AI researchers are taking alternative approaches.
THE ORIGINS OF RATIONAL THOUGHT
By the year 1965, programs existed that could solve any logical problem in the world, in theory. Much of this progress was due to the Greek philosopher Aristotle who was the first in his time to really come out with a correct way of thinking which became the basis for all reasoning processes. He knew that given a correct premise there was a way to always come to an accurate conclusion using logic. A group of Logicians in the 19th century developed notations for statements about objects and their relations to each other.
This approach still exists today, modern logicians are attempting to build programs based on logical thinking and rational thought. This logical notation method does come with many problems such as the fact that when knowledge is not 100% certain a program based on this approach will run into trouble during the reasoning process. Nevertheless, this tradition first taken by the logicians was a large step in the development of AI.
THE FOUNDATION OF ARTIFICIAL INTELLIGENCE
The real building blocks in the world of AI lie with philosophy. As stated before Aristotle (384-322 B.C.) was the first to formulate a set of laws governing rational thought. These laws in theory would allow someone to generate conclusions mechanically given a correct premise. Another philosopher named Ramon Lull(1232-1315) proposed that reasoning could actually be performed by machines. And even Leonardo da Vinci (1452-1519) designed a mechanical calculator based on many of these principles. Though he never got around to building it a recent reconstruction of his design proved it would have worked even back then.
Another philosopher named Rene Descartes (1596-1650) was a strong advocate of the power of reasoning and was one of the first to follow a philosophy called rationalism. Rationalists claim that there are significant ways in which our concepts and knowledge are gained independently of sense experience. They also believed in the power of intuition which was, and still is, a form of rational deduction.
Descartes also believed in dualism which claims that there is a part of the human mind that is outside nature (referring to a “soul”) that animals did not possess and in turn intelligent machine would not posses this either. All of this led to the rise of the empiricism movement which is an alternative to rationalism. Empiricism claimed that sense experience is the ultimate source of all our concepts and knowledge. The debate between rationalism and empiricism still rages on today and when these concepts first came out they became the true foundation of most AI research and progress.
EMPIRICISM VS RATIONALISM
The dispute between rationalism and empiricism concerns the extent to which we are dependent upon sense experience in our effort to gain knowledge. Rationalists generally have two parts to their argument. First they believe that there are cases where the content of our concepts or knowledge exceeds the information that sense experience can provide. Second, they construct accounts of how reason in some form or other provides that additional information about the world. Empiricists on the other hand, develop accounts of how experience provides the information that rationalists cite. They also attack the rationalists’ accounts of how reason is a source of concepts or knowledge.
Descartes and Aristotle were well known rationalists and the early empiricism movement began with Francis Bacon (1561-1626) and continued with John Locke(1632-1704) and many others who agreed with him. I will be describing these two differing philosophies in greater detail in a future blog post as there is far too much information to list all of it here. Also, it is an important piece of the foundation of the concepts that led us to all the progress in artificial intelligence that we have seen today.
THE BIRTH OF THE ARTIFICIAL INTELLIGENCE WORLD
Artificial intelligence was arguably born in the year 1956 when a prominent AI figure of his time named John McCarthy(1927-2011) received his PhD at Princeton and remained there for two years working as an instructor. McCarthy then moved to Dartmouth College which became the official birthplace of the field of AI. He, along with a team of researchers organized a two month workshop where they would attempt to study every aspect of learning or any other characteristic of intelligence that could possibly be simulated with a machine. They believed that if they spent the entire summer together that much progress would be made and many problems would be solved.
The result of this workshop was a reasoning program that was capable of thinking non-numerically and had proven many of the theories of the early philosophers true. However, this program had largely already been invented by two researchers prior to the workshop who just happened to be there. There weren’t any serious breakthroughs that took place in Dartmouth that summer but the main players in the field of AI had all been introduced to each other. The next 20 years of progress would come from this very group of people.
In the next part of this series I will go into the decades after the Dartmouth workshop and cover all the details that paved the way for modern AI technology and all the ways in which that workshop changed the world.