4

Why we need a new kind of computing Until the recent advent of what is popularly known as artificial intelligence (AI), the way computers compute hadn’t changed its underlying structure for nearly 70 years. Since their introduction in the 1940s, computers had required that data conform to strict rules and formats. Likewise, the questions we asked them had to be structured to follow the same well-defined logical path. These computer algorithms fail with today’s fastest growing and most pressing information challenges. They struggle understanding big data and are unable to handle the vast amount of unstructured text and images found in newspaper and magazine articles, blogs, social media and elsewhere—unstructured data accounts for 80 percent of the information humans create and consume today according to IBM Watson. We have arrived at a point where a new kind of computing is needed. We need computers that are able to process information similar to the human brain. Cognitive computing understands the problem and knows how to find the answer The thing that makes cognitive computing really different is that it isn’t a pre-programmed answer machine—it’s a learning machine with the ability to deduce the answer itself. In an article, “Demystifying artificial intelligence: What business leaders need to know about cognitive technologies,” (Deloitte University Press, November 4, 2014), authors David Schatsky, Craig Muraskin and Ragu Gurumurthy look at the business landscape for cognitive computing. They suggest that while the hype around Artificial Intelligence makes it sound “more like science fiction than it does an IT investment,” the technology is real and will be critical to business success. 4