Take a tour of the famous Turing test, as described by Expect Labs founder and CEO, Timothy Tuttle.
TRANSCRIPT:
I recently did a blog post about this long-standing debate in the AI community about what constitutes true intelligence. Is it is it just a collection of simple pattern matching or is there something fundamentally different about the nature of how the human brain works that can’t be simulated by today’s computer systems. I read this very interesting recent paper by Hector Levesque who’s a professor of computer science at the University of Toronto, and he discusses this issue and makes some very interesting proposals about how the way the computer scientists have been evaluating the intelligence systems over the past 50 years has perhaps been well, has been flawed, and there might be a better approach.
Historically, the way that computer scientists, at least as a thought exercise, have envisioned evaluating the intelligence of a computer system is using the well-known Turing Test, which is a theoretical test that was developed by Alan Turing where the way that you differentiate an intelligent computer program from a human is by having a computer in one room, and having a human being in another room and having a participant ask questions to each of these two entities through a teletype machine. The computer program would be considered to pass the Turing test, if after asking all the questions that the interviewer could think of, he was still unable to differentiate between the computer program and the human. That’s the Turing test.