The turing test isn't an effective way of determining a computer's intelligence because the test often limits what a person has the ability to say. Sometimes during these test people aren't allowed to try and stump these computers. To me a intelligent being is thing or someone I can have a conversation that makes progress. Maybe a more effective test would to have someone try to learn about history from one of these machines or get taught a math problem with Q&A during the session. It seems like all the chatbots out now are only good at relating what it is you have to say with other world in a huge database or changing around the syntax of worlds. Intelligence is the ability to grasp concepts or ideas that are not already known. If a computer isn't programed to know what an apple is it should be able to draw a picture of this in it's own "mind." It should be able to understand what an apple tastes like from the description a human gives it, or better yet, from another computer. It should also be able to some remotely related to the key aspect that distinct us from other animals: communication and understanding of ideas. The Turing Test can't weigh these kinds of distinctions. The difference between a computer having computer like comprehension and linguistic skills with human like inteligence is too broad of a gap for a simple test like the Turing Test. It should also be pointed out that this testing method was composed over 50 yeas ago. I think that now, in todays technological world, the test for AI intelligence should have a much higher standard to be met.
