Pim Haselager, philosopher-researcher at Radboud University and a specialist in intelligent robots, doesn’t expect that robots will ever take over the world because ‘we don’t know how to build a robot that has goals of its own’. Nevertheless, he does expect a revolution of sorts. In about thirty or forty years robots will, for example, make our food and clean the floor. They will actually help us to stay longer in our own homes when we are older.

But will you be able to trust your robot? Haselager demonstrated this issue with a video in which a cat, sitting on Roomba the vacuum robot, proceeds to attack a dog. ‘The people who built Roomba know this robot very well, but none of them had ever anticipated that a cat would use it as a tank.’ He emphasizes that the world is a dynamic place: there isn’t only the owner of the robot, but also many unpredictable environmental factors to take into account, such as kids and animals.

According to Haselager, we will have to give robots the capacity to assess their own capabilities. ‘In a nutshell, in order to build robots we can trust, we have to build robots that distrust themselves.’

Blog by Linda van der Pol