AlphaGoZero (1) was the first Robot which could learn to play “Go” without historical data. In Fact, AlphaGoZero did not play from “scratch” but was playing against AlphaGo, against on another robot.

Like the child who learns to play football by learning the codes and imitating the adult, the robot was learning by reinforcement: by the feedbacks of another robot.

It’s exactly like we learn with reward and punishment. Our entire educational system, – the stupid ones yes -, is based on this 2 tits of social education with teachers acting more or like the robot of AlphaGo.

We may imagine a first Robot A with the historical data “teaching” a Robot B ect.

After several robots, you will have a group of Robots with social norms. It means they will act in conformity of what the education system gave them.

All interaction between Robot, inside the group of Robots, will produce also new skills, social habits.

Emotions we feel come to the confrontation to the norms we have finally. It’s a very straight way, not really true but we could simplify it like that.

Then we can say that Robot(s) confronted with a different norm could begin to feel emotions – pain yes but emotion.

Going a little far away in a futurology scam, we could point that 2 groups of Robots may have different behaviours, habits and norms if they are separated for a long time.

When they will meet ? what would happen?

Ho Dear, it’s written. (2), aggressive methods are the best way to impose your norms like the conquistador with the Incas or the ones against Indians.

Am I sending bad vibes to Artificial Intelligence Fans? no, just pointing out that we will make new humans which will quite similar to the one occupying this planet.

(1) https://deepmind.com/blog/alphago-zero-learning-scratch/

(2) http://www.businessinsider.fr/us/google-deepmind-aggressive-gameplay-2017-2