Question: Can you also program a robot's feelings?

Giuseppe Cotugno answered on 10 Jan 2020:

Good question, depends how you define “feelings”. If a feeling is a prize or a punishment for doing something better or worse then we can say that we already have something like feelings in robotics. When a robot (or an AI system) has to learn things by trying them out for every attempt or move can be given a prize or a punishment (basically a number is summed or subtracted). This technique is used when robots are learning new things and is called reinforcement learning, and the robot / AI will try to get praised all the time by getting rewards.

If you have in mind more complex feeling, like love and hate, my first question to you would be “why do you need such a thing in a robot?”. A robot is basically a device which should make our life easier and more comfortable, so probably there is no use in it having feelings at all. I am aware, however, that some robots (like the Nao robot, a small humanoid) are used to support children with autism, so I guess those robots would benefit from showing some degree of feeling, but it still a machine used as medical device: its “feelings” are expressed to make people better not to create a friend.

With computers we can do a lot of things and probably one day we will be able to simulate feelings in machine up to a realistic level but the questions should always and always be “do we need this in a robot?” and “if a robot expresses realistic feelings and a human is tricked in believing them as genuine, which side effect can impact the life of the person?”.

To sum up, I doubt that we can create feelings in robots which are the same of a human’s but we might one day able to simulate them up to a realistic level, pretty much as some modern 3D movies look so realistic that is sometimes difficult to tell whether a picture of the movie is real or is 3D computer graphics