Turing Test Lab Post

Of the three tests, I found that the Alice worked the best because it responded to things very quickly, but all of them had their issues. Admittedly, this may be confirmation bias since I knew prior to clicking on the links that they were AI's rather than real people, but none of them were particularly convincing. At their core, all three machine tests had a large flaw: they don't act; they only react to whatever the human types in. If I ask about the weather, it can parse the sentence for the word “weather”, and spit out a generic response relating to hypothetical weather conditions. It can only react to key words. For instance, I asked Alice “Do you like Metallica?”. It responded “Not since they put the pressure on Napster.”. A reasonable response, but then I ask “Do you like Megadeth?”, being another metal band that most, if not all Metallica fans will also be familiar with. The response was “No, I don't care for it.”. Use of the pronoun “it” is not the way that a natural English speaker would respond. Most speakers would refer to a band as “they”, or if they're unfamiliar with them, ask “Who is Megadeth?”. The syntaxical errors these programs routinely make hurts their credibility. Singular and plural nouns are referred to by opposite pronouns, and the gendered pronouns he, she, and it are often used interchangeably. In addition, they're sensitive to typos. I asked Alice “Who is Barakc Obama?”, with the misspelling intentional on my part. It could not provide a correct answer. Asking it “Who is Barack Obama?” yielded the correct answer: the President of the United States. I don't believe that a real Turing test capable program will come out within the next 10 years. At its core, the problem with these tests is that they do not react in a realistic matter. But the test itself can be brought into question. Personally, I do not see it as a meaningful test of whether or not something is capable of thought. It is a test of how well a program can react and fool the test taker into thinking it is a real person. If there was to be something I would define as thought, it would have to react similar to something like the character Data from Star Trek: The Next Generation. While incapable of emotions like anger, Data is capable of learning and adapting himself to new experiences and thought processes. The programming is not set, as he accumulates knowledge and interacts, the more information he can retain, and is then capable of autonomous action without a human initiating the action or reaction.