Could you teach a computer to communicate like a human?

The hunt for computer sentience has been in progress for a long time. The first attempt at making an A.I (artificial Intelligence) was in the 1950’s. Since then, A.I’s have been well integrated into our daily lives.

 Video games make a good example of a ‘fake’ A.I. Many if not all games in modern times make use of an artificial intelligence, which control many of the objects and beings in the game world. These are merely programs, however, and have no intelligence beyond the coding entered by the programmer These A.Is are much more ‘skilled’ in games due to human error and the speed with which it can execute actions. Take the Automaton 2000 for example: This is an A.I made for the video game StarCraft II. To put it into perspective, an average RTS (Real Time Strategy) player will play at between 20 and 70 actions per minute. (APM) Professional StarCraft gamers will play at the speed of about 300 APM but, the Automaton 2000 plays can play at 15,000 APM!

However, that isn’t entirely what this article is about.

First, we must first identify what it means to communicate like a human. Thus, I shall introduce a list:

-          Language/Grammar

-          Tone

-          Emotion

-          Recognition

I mention this because a current computer can “talk” or communicate to you through messages on the screen. (Although it only knows you as “user name”) it cannot, however, humanely react to you in any way when speech is involved. For a computer to communicate like a human it must use language and tone appropriate to the situational context, as well as recognise the emotions of who it’s communicating with.

In order to teach a computer how to communicate in such a complex manner, it would take an even more complex program, a program that can teach it to recognise that you are having a bad day (making use of a webcam of course) and responding in the appropriate manner. On the subject of response, the program would have to teach it to recognise the situational context and what language would be appropriate. For example, if a ‘user’ had just lost a friend then it would be inappropriate for the computer to bring the subject up.

So, a program such as this would not only have to teach the computer each and every word in whatever language was in use at the location, but how the words relate to each other. It would also require some sort of coding that would make the computer put the words together on its own. Finally, the program would include coding emotion into the machine.

If a computer reached such a stage, it would probably be capable of doing many other things. When things get like that, I feel that hum

ans wouldn’t be very happy with so many computers doing their own thing. While I think that a robot rebellion is a bit far-fetched, I do feel that computers shouldn’t be integrated into our lives to such a degree.

So, it will probably be possible to teach a computer to communicate like a human in the future. Whether that’s a good idea is a different story.