The other day, I was reviewing a bunch of research papers from some computer scientists at Carnegie Mellon University. They had been working with advanced algorithms for chat bots, with the intention of being able to fool humans on Internet forums to assuming that the chatbot was actually another human that they were talking to. If they were able to do this, in essence they would have passed the Turing Test. Okay so let’s talk about this for second shall we?
First of all, should we condemn the AI scientists for attempting this maneuver on unsuspecting humans in the name of science? Not hardly, and perhaps if you were fooled by a chatbot, at least for a quick round of conversation how to build a chatbot anything got to technical and you realized that whoever or whatever you were talking to was wasting your time, that you might be upsetting. But, I’m sure you’d get over it, and think to yourself; hey that’s pretty cool, even if you aren’t part of the geek squad.
Now then, as I read these research reports, and watched the conversation of chatbots chatting each other up, I realized that the conversation degenerated quite quickly. Also, the chat bots were very quick to mimic human behavior, and start accusing the other individual in the conversation of misrepresentation, lying, or ignorance. It seems to me that we could grade the ability of these artificial intelligent chatbots, and those they are conversing with and rate them on a scale.
In fact, we might be able to tell the IQ level of the individual that the chatbot is talking to based on how quickly the conversation degrades itself, or how quickly the human decides to quit the conversation realizing they can’t win talking to this chatbot, even if they don’t realize it is a chatbot. In essence, it would be the human using mental game theory deciding that the conversation wasn’t worth the time or the energy to continue.
So my question might be; how long would that take for someone with an IQ level of 80, and IQ level of 100, an IQ level of 125, or a higher level IQ genius of 146 or higher to figure that out. Is it possible to use these chatbots in judging intelligence level of Internet forum users, bloggers, and those doing text messaging? It is my contention that it just might be. Therefore, I dare to ask the question, as I haven’t read anything about this yet. Surely, someone around the globe has posed the same question, and I would like to know if there is a potential answer. Indeed, I hope you will please consider all this and think on it.