IBM Chief: Computers Will Soon be Smarter Than Humans

Cognitive Computing

According to Inverse, an estimated 2.5 exabytes of data bombard us every day. That's a lot, considering the human brain can only store about 2.5 petabytes (or a million gigabytes), according to Northwester University psychology professor Paul Reber.  Overwhelmed yet? Well, over the next three decades, that amount of data we encounter daily is expected to rise to over four zetabytes. 

“The information being produced is far surpassing our ability to consume and make use of,” said Rob High, IBM Watson's vice president chief technology officer during a panel discussion at the American Institute of Aeronautics and Astronautics (AIAA) SciTech Conference 2017. He continued, “we have a very desperate need for cognitive computing.”

High believes that the current rate we produce and consume information requires mental prowess beyond what the mind of a single person — or even the minds of many persons — can comprehend.  Intelligent computer systems have to be a lot smarter in order to handle all of this information.

“If you’re a doctor and you’re trying to figure out the best way to treat your patient, you don’t have the time to go read the latest literature and apply that knowledge to that decision,” High explained. “In any scenario, we can’t possibly find and remember everything.”

Beyond Limitations

High believes that Cognitive computing is a possible solution to all this, and one that IBM is already working on. Currently, IBM is working on machines that would focus on human experiences, capable of helping, teaching, and inspiring. As a result, these computers will be smarter than humans – so we need to be prepared. Researchers are even investing in studying the ethics of AI to ensure it benefits humanity.

AI systems have been slowly getting better and better. In some cases, they've even outperformed their human counterparts, like in that historic game of Go. More recently, this same AI — Google's DeepMind — that defeated the world champion of Go also started killing it in an online version of the game. Google's DeepMind has also managed to outperform professional lip readers. Other examples include an AI version of J. S. Bach that can compose its own cantatas and an AI that can learn on its own.

This is all good news, according to High. We need AI systems that can assist us in what we do, particularly in processing all the information we are exposed to on a regular basis — data that's bound to even grow exponentially in the next couple of years.

“Smartphones are just the tip of the iceberg,” High said. “Human intelligence has its limitations and artificial intelligence is going to evolve in a lot of ways that won’t be similar to human intelligence. But, I think they will work best in the presence of humans.”