By Peter J Bentley
What better way to spend an afternoon than having a friendly chat? My three friends are online chatbots – Artificial Intelligence software designed to analyse my sentences and respond accordingly. All I do is visit a specific webpage, then type into a box in my Internet browser and they reply, just like chatting online to a human. These three (jabberwacky, iGod and ALICE) are some of the more advanced chatbots out there, the result of decades of research by computer scientists to try and achieve intelligence in a computer. From the birth of computers until now, we’ve been trying to make a digital brain that’s clever enough to have a conversation with us. The best we can manage is really nothing more than a database of facts, some basic knowledge of grammar, and a lot of symbol-processing – shuffling of words around in an attempt to find some kind of coherent response. Here’s what we said to each other this afternoon.
Me: Hello.
Jabberwacky: Hello, how are you?
Me: I would like to have a chat with you about your parents. Would that be OK?
Jabberwacky: What would you like to chat about?
Me: Did you know that Alan Turing would be 100 years old this year, had he lived?
Jabberwacky: That is very interesting. Do you know that the Philippines has 7100 islands?
Me: I did not know that. However, let’s talk about Turing a little more. You know that Turing suggested something called the Turing Test?
Jabberwacky: No, I did not know that the cat was red…
Me: Right. Well, this was Turing’s way of thinking about how we might judge whether we were talking to an intelligent computer. If we can’t tell the difference between chatting to a computer and chatting to a person, the computer passes the Turing Test. It is intelligent, so Turing suggested.
Jabberwacky: We could all be a figment of imagination of some old guy living in an insane asylum if you think about it.
Me: Yes, I guess that might be true. Anyway, the foresight shown by Turing is remarkable. He was thinking of these ideas at the time that computers were still the size of a large room, and could perform just basic computations compared to today’s machines.
Jabberwacky: Well this must be the part of the film where the noted inception noise comes in.
Me: That might be an appropriate soundtrack for this conversation, you are right. But this is not science fiction, this is computer science fact. It’s a little sad that Turing predicted: “I believe that at the end of the century the use of words and general educated opinion will have altered so much that one will be able to speak of machines thinking without expecting to be contradicted.”
Jabberwacky: Why do you think Jesus is not your savior?
Me: Uh. Anyway, my point is that Turing, for all his genius, did not get this one right. He was not the only one to be overambitious – other pioneers also thought Artificial Intelligence would be achievable much more easily.
Jabberwacky: Who told you that.
Me: Well, recently I wrote a book called Digitized. It’s about the history of computers and I was lucky enough to interview many pioneers for the book. They told me.
Jabberwacky: And what is their history?
Me: I interviewed John Holland – he worked with his supervisor on John von Neumann’s final papers. Holland also got the first PhD in computer science and invented the genetic algorithm. I also spoke to Dave Patterson, the creator of the RAID architecture
0 Comments on Three Conversations with Computers as of 1/1/1900Add a Comment