Suppose you conducted an experiment where Speaker A and B are talking to you on separate phones.
You initiate a conversation with A and B – posing the exact same questions to each speaker.
As the conversation progresses to its conclusion, you are asked to determine which speaker, A or B is a computer and which is a human.
The 65-year-old Turing Test is successfully passed if a computer is mistaken for a human more than 30% of the time during a series of five-minute keyboard conversations.
This test was posed as the imitation game by Alan Turing – a pioneer in the Computer Science, Cryptography, Biology, Mathematics and Chemistry.
After WWII, Turing focused his attention to developing computers that could learn and think. Alas, no such device was ever constructed. Turing did publish seminal papers on the topic Computing Machinery and Intelligence (1950)
Turing proposed the Imitation Game as a more succinct way of asking “Can machines think?”
In response to the term “Machine Intelligence”, with two questions – “What is a Machine?” and “What is thinking”? He sort of skates around the question by saying a thinking computer would respond in the same way to a question as would a human without a definitive yes or no.
The questions are as well an illustration of infinite regression where the question “Can computers think?” is inverted into two questions –“What is a Machine” and “What is a question?”.
One question to ask any computer is the following question asked by Monty Python in the World Forum/Communist quiz sketch – “In which year Coventry City FC last won the FA Cup?” It is a trick question – Coventry City has never won the FA Cup.
CleverBot (CleverBot) claims to have passed this test. When asked the Monty Python question, the system answered “I don’t follow Sports. Use Google”.
When I was in University in the late 1970’s/early 1980’s, there was a game called Doctor which spoofed the imitation game. It would ask “What seems to be the problem ?” You would type in I have a chest cold. The application would restructure the response into a question and gather additional information about the “Patient” as the conversation progressed. So it would ask “Does it bother you that you have a chest cold? Using the principle of infinite regression, Doctor would endlessly transform a statement “I have a cold” into a question. This format of discourse can carry on indefinitely.
The computer gives the allusion of intelligence but is doing no more than being an electronic parrot.
YouTube videos of two chat bots arguing with each other using infinite regression to structure the discourse where one bot makes a statement and asks the other bot a question or insults the other bot when there is nothing to say. (Bots Arguing)
I conducted research in AI in the early 1980’s. My hypothesis was that learning is about modifying the model of information we maintain about the world or a specific topic to a new form. The model is communicated for example, in our thoughts, speech, writing and music.
The state of the AI art can be characterized as slow going with spurts of intense research and then silence. One spurt of development was in the area of expert systems where a long list of if then else rules and corresponding database of facts. These systems are nothing more than rules engines used for example in call centres to figure out how to up sell the caller. For example, an online sales system detects that you’ve added a red widget to your cart. It could consult a rules engine and find out that the shopper is 52% likely to add a brown widget to the cart. The shopping system would generate an offer for a brown widget using a rule for up selling shoppers. The database and rules engine need to be continually updated for the system to adapt to a changing environments.
2017 can be characterized as AI is again on the ascendancy in the form of self driving cars, semi trucks and ships.