Leadership in the age of AI
From manufacturing and city infrastructure to the human body itself, the digital revolution will reach into every corner of our experience. Only with the help of artificial intelligence will we be able to make sense of these increasingly complex, data-rich environments. I-CIO sat down with Fujitsu’s chief evangelist, Iwao Nakayama, to discuss the implications for business, society and the role of IT leaders.
• Fujitsu Forum Tokyo 2018 Speaker
Some people immediately connect AI to technological singularity and start talking about machines evolving beyond human intelligence or about nano-devices being injected into the human body to access biological information or digitalize brain activities. Today, this is still within the realm of science fiction. Ray Kurzweil, the world-renowned futurist, predicts that singularity will occur around 2045 and be recognized as a change that redefines humanity. But we haven’t arrived there yet.
That said, I’m sure that AI will trigger paradigm shifts in many industries in the near future, because no industry – from agriculture to engineering to service – will escape digitalization and the abundance of data this produces. And companies must implement AI to take advantage of, and extract value from, data.
Let me give an example to help illustrate it. If you were in New York City back in 1900, you would have seen countless horse-drawn carriages on streets like Fifth Avenue. But by 1913, you couldn’t see them any more – all you’d see was a parade of Ford Model Ts. The change – the paradigm shift in mobility – happened during a period of only 13 years.
I want to emphasize how short a time that was. The Ford Model T was first shipped in 1908, so the change took only five years. Paradigm shifts often happen this way, transforming everything rapidly.
Automobiles not only replaced horses on the streets, but also ushered in a whole new ecosystem to serve the motoring public, such as gas stations, traffic lights and driver’s license bureaus. I think AI is a technology with the potential for this type of transformation and we are on the eve of an AI-triggered paradigm shift.
Currently, Fujitsu is working with customers on about 800 AI-related projects. One of the initiatives, for example, is a collaboration with Hiroshima University to build an AI-based technology that enhances doctors’ ability to rapidly retrieve relevant cases from a computed tomography (CT) scan database. Following the implementation of the system, which is being used to diagnose diseases such as lung cancer, doctors are able to do the same task six times faster.
In manufacturing environments, some factories have built anomaly detection systems that instantly identify faulty products, using the image recognition capabilities of AI. For accurate detection, the factories employ the deep-learning capabilities of AI to teach monitoring equipment to recognize what normal products look like. This system is also attractive to the food industry because it can detect product tampering in packaged foods.
AI has also been used for several years in call centers, where operators’ interactions with customers constantly generate speech and text data that AI can analyze and learn from. AI is then able to find the most useful information at the moment it is needed out of trillions of bytes of data. Of course, not all AI implementations are successful. Some face fundamental problems.
You can’t expect good results if you start the implementation without articulating the specific areas for which you want to use AI and also without closely studying the types of data your organization creates and stores.
AI is involved in technologies such as machine learning, deep learning, image recognition and speech recognition, but I think it is essentially a technology to optimize the use of data. Therefore, you must have specialists in the field – namely, data scientists and data analysts. Without these resources or skill sets, your projects are likely to hit a wall.
At some call centers, for example, management teams learnt that most of the text data they owned was not really usable for AI analysis only after they found it was full of abbreviations and internal jargon. So it’s important to understand AI technologies and your own data before you set out on projects.
Although many people in business have already recognized the importance of the CIO, the role is going to become even more important for a company’s success.
The CIO needs to take into consideration not only the company’s customers and markets but also its workers when he or she develops the technology strategy of the company. A good example is Goldman Sachs in the US, which is said to have reduced the number of traders in one of its offices from 600 to just two after implementing AI. Fortunately, the 598 were relocated to other roles. A CIO today should be aware of the impact technology will have on the workplace.
In short, he or she has to drive the deployment of new technologies across the company, consistent with the strategy of the CEO, while being mindful of the likely effect of the changes on workers.
Businesses implementing AI solutions will require higher computing capability than ever before. Traditional computers could take too much time in solving today’s complex problems. To give a practical alternative to those who are waiting for quantum computers, Fujitsu has developed a futuristic computing technology with a new architecture, Digital Annealer, which brings the potential of quantum computing but delivered through general-purpose computing technology. Digital Annealer is able to solve a combinatorial optimization problem in a matter of seconds, a problem that would take a supercomputer hundreds of millions of years. As that illustrates, we continue to help our customers turn the challenges of paradigm shifts into opportunities for their organizations.