Table Speech


How Will AI Change the Society?
--- The Present and Future of Watson ---

April 12, 2017

Ph.D. Kazushi Kuse
Vice President, Chief Technology Officer,
IBM Japan


 The history of intelligent computers started with tabulating machines, shifting to the era of programmable systems and then to the so-called “cognitive computing” era that gives computers the ability to learn. IBM developed the first cognitive system Watson, which can perform complex analytics on huge volumes of unstructured data without requiring human assistance. At IBM, when we say AI it means “Augmented Intelligence” not “Artificial Intelligence”. We are engaged in the development of a system which “augments human intelligence to improve our performance”. AI is not meant to outsmart humans but to work in partnership to assist us.

 The Watson project was launched in 2006 by team members with cutting-edge expertise from fundamental research laboratories in 8 countries, including Japanese researchers specialized in natural language processing. In February 2011, Watson made a sensational win over two accomplished champions of Jeopardy, an American long-running TV quiz show. Watson has read, analyzed, learned and loaded information worth 200 million pages on diversified subjects. It is capable of giving answers by interpreting what is asked, understanding the context and extracting logical responses from multiple options. Right after Jeopardy was aired live, we received inquiries from many companies, universities and hospitals on practical application of Watson for business and medical practice. We engaged in joint verification initiatives with multiple companies and hospitals over four years and established the IBM Watson Group in 2014 to promote the development and commercialization of cognitive innovations.

 Watson is most widely used in medical settings, providing assistance to physicians in diagnosis and identifying the most effective treatment, based on a massive volume of latest information and medical data acquired. Competent physicians are usually very busy and have only five hours a month on average to study the latest developments of medicine, treatment and new cases. Watson has been providing efficient support at medical institutions, including Memorial Sloan-Kettering Cancer Center, MD Anderson Cancer Center and Cleveland Clinic. Here in Japan, a hospital affiliated to the University of Tokyo’s Institute of Medical Science has introduced Watson loaded with 25 million medical papers and 15 million pharmaceutical data. Watson saved life of a female patient in her 60s by providing diagnostic assistance for a rare type of leukemia and identifying the most effective therapy. The patient was initially diagnosed with acute myeloid leukemia and underwent chemotherapy at the hospital over several months. Unfortunately, however, her condition got worse and the Institute turned to Watson to detect the real cause of her illness. Researchers supplied the patient’s genetic data and Watson cross-checked it with the database. Based on Watson’s analysis, the physicians detected a rare secondary leukemia and changed the patient’s therapy plan, after which her condition improved significantly and she could be discharged from the hospital. Watson succeeded in providing the data physicians needed in an extremely speedy fashion. This example showcases an ideal way of collaboration between humans and computers in overcoming challenging issues like improving clinical therapy.

 Chef Watson is another cognitive computing application that can enhance human creativity. IBM researchers brought Watson to a culinary school in New York for nearly one year to analyze 30,000 recipes and read up on the chemical composition and health effect of numerous ingredients. By combining data and detecting certain patterns, Chef Watson suggests novel combinations of ingredients that help professional chefs come up with new recipes. Here in Japan, we also present delicious and healthy recipes for home chefs through a cooking website Cookpad. Watson is expanding its scope of application around the world in multiple sectors, including education, finance, media, manufacturing and logistics. IBM implements the Watson IoT Partner Ecosystem program to make the Watson operation simpler and more useful to wide-ranging partners. Currently, we have over 500 partners, including universities and venture companies. “Cognitoy” is a small educational dinosaur toy powered by IBM Watson. It can have interactive dialogue and engage children in educational play. A self-driving electric-powered minibus by Local Motors is another example, currently in pilot operation at Washington D.C., Miami and Las Vegas. Watson plays the role of a conductor and engages in conversation with passengers.

 Started as an IT system consisting of one big hardware and software, Watson has evolved into a multiple application system that allows combination of different functions for logical analyses. Future development of cognitive computing and AI will require high-performance computers that consume large amounts of electricity. Watson that won the quiz show consumed 200kW whereas human brain consumes merely 20W. IBM is currently working on the development of high-performance neurosynaptic computer chips with extremely low electricity consumption, about one thousandth of conventional CPU. IBM is also conducting research and development on quantum computers, among many other companies. I believe driving the study and adoption of quantum computing worldwide will be the key for the future.