Illustrations for an article about MIT experiments with a new method of cooperation between people and robots (source https://www.csail.mit.edu/news/how-control-robots-brainwaves-and-hand-gestures). In order for a robot to understand whether it is doing things right it needs to determine a man's reaction to its actions - man's gestures and brainwaves changes.

The controlling person may get distracted by another person. In this case the robot may start doing harmful things, without being aware of that. My first illustration depics this idea.

Then I developed this idea that some evil person may encourage the robot to do harm and destroy everything. (Red text: do not skip)

Actually watching the robot's work and analyzing its actions is similar to raising one's own baby. Ii the end it adopts the worldview and values of its parents.