Will cognitive computing, which enables us to see clearly through the mass of data surrounding us, change the face of the world?
IBM’s supercomputer Watson, which in 2011 became a superstar following its victory on television quiz show Jeopardy! over Ken Jennings and Brad Rutters, two of the game’s best players, is back in the spotlight. Watson has now hit the headlines again due to recent advances in artificial intelligence (AI), illustrated in particular by the performance of another thinking machine, this time developed by Google. In March, to everyone’s amazement, the AlphaGo programme from DeepMind – run by Google's AI lab in London – managed to beat one of the best Go players in the world, without any handicap. The potential of AI is today on everyone’s lips, generating enthusiasm and fear in equal measure.
However, focusing exclusively on the highly publicised feats achieved by AI makes it very difficult to discern the future that these impressive machines are now shaping. In fact IBM intends to use the technology that brought Watson to TV stardom to actually “change the face of the world,” as John Kelly, IBM Senior Vice President for Solutions Portfolio and Research, put it at the EmTech Digital event on artificial intelligence hosted by MIT Technology Review, which took place in San Francisco in May.
Making the invisible visible
So what exactly is Watson? “People generally believe that Watson is an attempt to replicate what the human brain does. That’s not at all what it’s about,” clarified John Kelly, explaining: “Our objective is to help the human brain manage the massive amounts of data it’s confronted with.” He reminded the audience that the vast majority of data generated today cannot be used: “90% of data available in the world has been generated in the last two years. But human intelligence is not growing at the same speed as the quantity of data. So 80% of data generated today is unreadable and unusable, which is a huge waste. We can only see a small part of reality.”
This ‘dark’, invisible data can be compared to the dark matter well-known to astrophysicists – i.e. matter that cannot be observed directly, but whose effects on the universe can be observed. In a similar way, the ‘dark’ data that is all around us has a direct impact on our existence, but for the moment we are unable to use it. It is IBM’s ambition to use Watson-style computing to change that.
Similar to dark matter, ‘dark’, invisible data has an impact although we cannot currently make sense of it.
The third computer revolution
John Kelly argues that we are now seeing the dawn of a third revolution in the computer world. “The first era of computing was between 1900 and 1940. Computers were then straightforward tabulating machines, largely mechanical. People entered data manually and then let the machine process it, basically automating certain tasks. From the 1950s on, we entered the next era, with programmable systems. Computers started to have enough memory to run themselves without any outside intervention. But those computers cannot process the exponentially-increasing mass of data being generated today.”
John Kelly explains that we are now witnessing the development of a third type of computer, which is more suited to processing vast amounts of data. This cognitive computer “will be capable of learning on a huge scale, processing immense quantities of data, drawing conclusions in relation to an objective set in advance, and communicating the conclusions to human beings so that they can act accordingly.” Not an artificial human brain then, but a kind of super-brain, which will focus on tasks that are far too great for human intelligence, opening up a whole range of new opportunities.
Impact across many sectors
Cognitive computing will transform all sectors of the economy. The oil industry, for example, today generates astronomical quantities of data, most of which is simply wasted. “These days, oil platforms are installed with tens of thousands of sensors,” Kelly points out. If the operator can collect and process all the data they generate, that will help to radically improve efficiency and bring costs down. IBM could for example work with [French multinational integrated oil and gas company] Total in this field.
These developments will also have a major impact on the retail business. Every day, millions of people give their opinions on goods or services on the social networks. This potentially provides brands with huge amounts of precious information but it is still largely under-exploited. Three-way collaboration between major retailers, IBM and Facebook might well be on the horizon.
This ‘third computer revolution’ will also help to create smarter cities. For instance, processing traffic data will help optimise transportation networks, road signal systems, and the way cities are laid out in general. It will also serve to accelerate the advent of self-driving cars. “By 2020, 75% of all cars around the world will be connected, producing huge amounts of data in real time. In order to become autonomous, these cars will have to be able to take decisions, reacting fast to their environment and anticipating the behaviour of other drivers, by gathering, processing and incorporating large quantities of data, fast,” Kelly predicted.
Combating crime would take a giant step forward if the pictures taken by surveillance cameras could be used to better effect. In New-York alone, cameras generate 520 terabytes of data every day, but for the most part this is raw and unstructured.
John Kelly also reckons that computer security will look quite different. “The important thing will no longer be to develop firewalls and antivirus programmes but rather to carrty out in-depth analysis of Internet user behaviour and operating systems, anticipate anomalies and act accordingly,” he told the EmTech Digital audience.
In addition, weather patterns, which impact a whole range of industries, could also be worked out more accurately: “We’ll soon be able to predict variations in retail activity from weather patterns,” Kelly believes.
Not least, Watson-type cognitive capabilities could also be used to improve computers’ language ability and help refine interactions with human beings. Along these lines, US motor vehicle manufacturing company Local Motors is now partnering with IBM to build the Olli bus, an autonomous shuttle bus that can converse with passengers and answer their questions. This system could be used by all self-driving car manufacturers. It is also likely to add substantially to the capabilities of virtual assistants.
Gauging the weather effects on socio-economic activity is also part of Watson’s skillset.
The cognitive computer will also be able to adapt to each individual person, moving from an ‘off the peg’ to a ‘tailor-made’ approach. This is likely to bring radical transformation to two areas, on which IBM has decided to focus, education and health. Rashmy Chatterjee, IBM’s Chief Marketing Officer for North America, with whom L’Atelier spoke at the TiEcon 2016 event in Santa Clara, California in May, predicts that Watson will be able to help improve education by adapting to the specific cognitive features of each student: “In the education world today, all the children in the same class get the same teaching. Think of the advantages of education adapted to each individual, with a curriculum and teaching methods tailored to his/her profile”. Foresees John Kelly: “A student with a good visual memory would be given visual content. The teacher would gather data on the way his/her students learn best and act accordingly.”
Watson, medical assistant?
In the medical field, a more personalised approach could save many lives. “In the United States, 30 to 40% of the $3.5 trillion spent each year on health care is wasted because of a lack of information,” claims John Kelly. “On average, every person generates a million gigabytes of data on his/her health during his/her life which could reduce the number of medical errors responsible for many deaths.” He told the audience that cognitive computing could inter alia help improve information-gathering from medical imaging: “Today radiologists look at thousands of images – MRIs, X-rays, scans – every day. Drawing on a large medical imaging database, Watson could extract the images containing important information, look at these together with the patient’s medical file and similar cases from the past, and so provide the doctor with more easily processable information, which will help to improve decision-making.”
Last but not least, working with a patient’s genome, Watson could identify the significance of each mutation and recommend specific treatment. And more generally, with the same aim, automate the most laborious and repetitive tasks so as to help improve human decision-making. As Thomas J. Watson Jr., IBM’s second CEO and son of the Thomas J. Watson after whom the ‘supercomputer’ is named, famously put it: “Computing will never rob man of his initiative or replace the need for creative thinking. By freeing man from the more menial or repetitive forms of thinking, computers will actually increase the opportunities for the full use of human reason.”