Above: Image © istockphoto.com/kapley
Computers and gadgets seem to be getting smarter every day. In light of these advances, what does the future of computing hold? The consensus seems to be artificial intelligence (AI). But what exactly is artificial intelligence? Does it really involve Terminators taking over the world? Hopefully not! In fact, Hollywood movies aside, the future of computing looks very bright.
How we’re teaching computers to understand pictures
Fei Fei Li, TED Talks
AI is commonly defined as giving human-like intelligence to a machine. Scientists have actually been working on this challenge for decades. In fact, a lot of the technology they have developed over the years has already made its way into your hands. However, they are still a far cry from developing a machine as smart as a human. The advances made so far pale in comparison to the long-term potential of truly intelligent machines.
One big challenge for AI researchers is sensory perception, especially vision. To understand why it’s so hard to give human-like vision to a machine, take a look at a cluttered desk. Chances are, most of the objects you see are partially hidden by other objects. But thanks to your knowledge of the world and the way your brain works, you have absolutely no trouble figuring out what’s what. I could also show you profile (sideways) picture of your friend’s face, and you will immediately know who it is. However, tasks like these are extremely difficult for computers.
Did you know? Watson, IBM’s artificially intelligent computer system, is being used to help find a cure for cancer.Despite these challenges, computer scientists have made great progress toward machines that can see like people. For example, Facebook recently launched a feature that recognizes your face in pictures, using your existing photos as a template. It uses information such as the distance between your eyes, nose, and ears to find you in new pictures.
Another challenge is building a machine that can understand human language. Apple’s virtual assistant Siri does mimic human dialogue. Siri compares the questions you ask her to a memory bank full of words and phrases. She then uses statistical analysis to determine how to interpret and answer your request. But how would Siri fare with an expression she was not programmed to comprehend? Could she understand a sentence like “The pink giraffe climbed on top of the Eiffel tower”? I bet you could!
What happens when our computers get smarter than we are?
Nick Bostrom, TED Talks
Part of the problem is the sheer complexity of the human brain. A computer can process information at blazing speed, but its way of thinking is very linear. It simply reads through strings of 0s and 1s, known as bits, one after another. With humans, things are a bit more complex. Your brain contains about 100 billion neurons (cells that transmit information) and about 1000 trillion synapses (connections between neurons). More importantly, brain cells and different parts of the brain affect each other as they work. All this makes for an extremely complicated network of information. And scientists are still far from fully understanding it.
But it’s important to keep pursuing AI. The potential benefits of intelligent machines go far beyond playing video games on your smartphone. For example, they could build simulations of scientific experiments. A doctor could create a virtual version of you to see if you might react badly to a certain drug. Or they could test a hundred different drugs to figure out which one will treat you best.
Intelligent machines could also help process huge amounts of information. As the amount of available information constantly expands, it is increasingly difficult to make sense of it all. But AI could allow scientists to feed a ton of information from different fields into a single intelligent machine. The computer could then look for patterns and possibly find surprising answers to incredibly complex questions. It might even reveal the secrets of time travel or the meaning of life!
Did you know? For most of history, human knowledge doubled about every 1500 years. Today, human knowledge doubles every year. IBM believes it could double every twelve hours in the not-so-distant future.Still, the potential of AI is not entirely positive. For example, philosopher and technologist Nick Bostrom thinks the fear of Terminators taking over the world is a somewhat valid concern. However, he believes that catastrophic scenarios can be avoided through careful research and planning.
In 2005, the Association for Advancement of Artificial Intelligence held a conference to discuss the growing need to study machine ethics. This new field of study looks for ways of making sure machines behave ethically toward humans. Of course, this also requires more research into how people make ethical decisions and what it really means to “do the right thing”.
Indeed, the future of AI holds great promise along with great challenges. As long as researchers develop intelligent machines in a responsible manner, the potential benefits could exceed your wildest imagination. So if you’re planning on working in the field of AI, take a page from Spider-Man and remember that “with great power comes great responsibility”.
Magazine article on the challenge of building ethical robots:
Podcast on significant developments in the field of artificial intelligence: