By SCOTT DEWING
Published: November 2006

You’ve probably said that to your computer too. And you’re right—computers are stupid. They can’t think. They don’t have any commonsense. “Oh, like a teenager,” you might say. Yeah, I suppose kind of like that with the big difference being that computers are very good at doing exactly what they are told to do, which cannot, for the most part, be said of teenagers.

Why is it then that our computers are, as we say, “stupid”? Why can’t they think like we do?

This was the very question that a young MIT doctoral student named Push Singh asked himself a couple of years ago.

How is it that we can write software that can do such complex things as design airplane engines, but still we cannot build machines that can look at a typical photograph and describe what is in it, or that can read the simplest children’s story and answer questions about it? We have been able to write programs that exceed the capabilities of experts, yet we have not been able to write programs that match the level of a three year old child at recognizing objects, understanding sentences, or drawing the simplest conclusions about ordinary life. Why is it that we can’t seem to make computers that can think about the world as any person can?

Push Singh asked simple questions but had the brilliance to begin tackling the complex answers. Tragically, he died in February of this year and the field of computer science and artificial intelligence lost a bright shining star to the darkness of an apparent suicide. Before he died, however, Singh began blazing a trail toward the answer to his question. He was convinced that the crux of the problem was that computers had no commonsense knowledge.

The real problem is that computers do not know anything about us! Our machines lack commonsense…Computers do not know what we look like, how we typically behave, or what we are capable of. They do not know anything about the patterns of people’s lives…They know nothing of our hopes and fears, the things we like and the things we loathe, or the feelings and emotions that motivate and underlie everything we do.

Singh was involved in MIT Media Lab’s Commonsense Computing Initiative. According to the MIT website, the purpose of the project was, “to give computers and other modern devices commonsense, the capacity to understand and reason about the world as intimately as people do.” The MIT postdoctoral students and faculty involved in the project believed that by giving machines commonsense, the machines would finally be able to understand people, our goals and typical problems. They believed that machines would then be able to better assist us in solving complex problems or, at least, help us to come to terms with those problems.

To build a computer with commonsense, one must first define and collect all the commonsense data there is out there. For example: touching something hot will burn you; if you stay underwater too long, you’ll drown; every person is younger than his or her mother; if you grab a knife by the blade, you’re likely to get cut. And so on. As you can see, there is a lot of commonsense knowledge. Though there is no conclusive number, several estimates have agreed that the number of pieces of commonsense knowledge is somewhere in the hundreds of millions.

Collecting all that commonsense knowledge would be a daunting task. Undaunted, Singh’s answer was to create a distributed human project called Open Mind Commonsense that would collect everyday commonsense data from thousands of human participants. “Computers today are just plain dumb!” wrote Singh on the homepage of the Open Mind Commonsense website. “[This] project is an attempt to make computers smarter by making it easy and fun for people all over the world to work together to give computers the millions of pieces of ordinary knowledge that constitute commonsense…This repository of knowledge will enable us to create more intelligent and sociable software, build human-like robots, and better understand the structure our own minds.”

One thing that Singh and his colleagues at the MIT Media Lab did with this commonsense data was build a search engine application that used commonsense knowledge to reason the true goal behind a user’s search query. For example, if a user typed “my cat is sick” into the search engine, the system would reason the following based on its collection of commonsense knowledge:

People care about their pets:

• and people want their pets to be healthy,
• and my cat is my pet,
• and I want my cat to be healthy,
• and a veterinarian heals sick pets,
• and a veterinarian makes sick pets healthy,
• and I want to call a veterinarian,
• and a veterinarian is a local service

Therefore: search for a veterinarian in the user’s area

While current search engines such as Google have the ability to provide you with the addresses and phone numbers of local veterinarians, none of them have the ability to reason what the true intention of a user’s search is. Imagine how much more powerful, informative and useful such a system would be.

What I dream of is a computer system that retains all of its capability to store and manipulate information as well as perform complex calculations at high speed, but has the added ability to learn and adapt as efficiently as the human brain. In short, I dream of a machine that is a thinking machine, a machine that matures, a machine that gains wisdom, a machine that knows everything there is to know, has commonsense and never dies.

But when I awake from the dream of that machine and what it would be capable of, I pause and wonder if it is a good idea to pursue the creation of such a machine. My commonsense tells me it’s not and that it is perhaps better if our machines remain stupider than we are.