Group 4 Language Blog

I find that people are not typically as unnerved by talking machines that are traditionally thought of to be silent as they are by usual human interfaces being replaced by computers. Talking cars and ATMs are still somewhat of a novelty, but are found extremely useful and not too offensive. These voices are becoming more and more human-like in use, so why not make them as human like as possible and include speech disfluencies, and at the very least they should be able to understand ours. Most people (sadly including me) give electronic devices, such as a computer, a human name because they have become such an important part of their lives. (Fredrick and I have a very strong relationship, lol, we're going on a year now and considering kids soon.) The difficulty with the use of talking computers seems to revolve around more sensitive subjects, such as banking or as eluded to in class a computer based psychiatrist that is able to listen and provide feedback. The future will no doubt bring up many questions like this... and I am very interested to see how we will handle them. There is certainly a limit to non-human interaction, we just have to find the point at which we comfortable with these new technologies.

I think that artificial intelligence raises a lot of questions. I’m not sure however, that I agree with the term “teach”. I would assume that most computers have to be programmed to respond in a certain way. I guess if a machine were able to truly communicate it would have the capacity to learn and acquire new information. If a computer could do that, it raises some very interesting questions. Should computers be able to "think" like humans? And if they can think, are they then conscious beings?

Beyond whether we actually want computers to be able to comprehend, the idea of computers communicating in a personable way is another interesting topic. For a computer to have a personable conversation, it must be able to insert uhs and ums. Without these natural disfluencies, a human would probably feel uncomfortable and not trust the computer. I do think however, that a computer could be programmed to add in uhs and ums at appropriate times. Current research suggests some patterns as to when we use uhs and ums, such as before new ideas are introduced. With the knowledge about these patterns, a computer could be programmed to insert them in the proper locations, making the human listener feel more comfortable.

However, a computer’s comprehension of human speech, I feel would be a much more difficult area to master. Even if the computer could somehow “ignore” the ums and uhs, there are so many other mistakes that would be impossible to follow. Things such as spoonerisms and sentence restarting are so sporadic that it would be very difficult to program a computer to correctly interpret these sentences. Therefore, in order for a computer to understand humans, we’ll need to speak in computer talk (kind of like how there’s a baby talk). This straight forward speech would seem to be the only way a computer could process our sentences.