New-Tech Europe | February 2019

AI systems will become our workmates. Workmates we understand and trust.

Pieter Ballon, Imec

AI is an evolution, not a revolution Science fiction films featuring robots or intelligent machines in the lead- ing roles (such as Blade Runner, Real Humans, Westworld, etc.) have caused us to look at a future with AI with some trepidation. But it won’t happen overnight and we will also have time to adjust ourselves to the idea and to control AI systems where necessary so that it becomes a gradual evolution, not a sudden revolution. But it is definitely an evolu-tion that is already underway. Harvard professor, Michael Porter, sets out four stages that mark the way toward smart objects and systems. Stage one is ‘Monitoring’: by using sensors, a smart product will be aware of its own situation and the world around it. An example of this is the Medtronic glucose

Artificial intelligence, or AI is all the rage again. Some people – most of them technologists – are looking at AI as a way to resolve some of the problems we face. But others are afraid of it. How can we make sure that AI systems – such as robots – will really help us and not take over the world and snatch our jobs away from us? Pieter Ballon, director of SMIT (an imec research group at VUB), emphasizes that engineers and social scientists need to work together on AI, because artificial intelli-gence is a technological innovation that will undoubtedly cause signifi- cant economic disruption and social changes.

meter, which uses a subcutaneous sensor to measure a patient’s blood- sugar level, alerting the patient 30 minutes before that level reaches an alarming status. Stage two is ‘Control’: thanks to its in-built algorithms, the product will then carry out an action based on the readings or measurements it has taken. For example, if a smart camera detects a car with a specific number plate, the gate will open. Systems then evolve towards the stage of ‘Optimization’. Basing itself on all the data that the system collects while it is operating, in-built algorithms can carry out analyses to determine the best way of work-ing. It’s as though the system ‘learns’ to work more efficiently. An ex-ample of this are wind turbines that are able to adjust the position of their vanes each time the wind changes direction so that they can cap-

20 l New-Tech Magazine Europe