There is a difference between people and machines: empathy.  I don't think they have successfully programmed empathy yet.  But who knows, maybe it will be achievable some day.

Psychological research has indicated that immoral acts, in the minds of humans, are innately considered 'impossible' tasks. Only when we take the effort to overcome the idea that an immoral act is both possible, and desirable, can we actually contemplate its occurrence and enaction.

A similar method could easily be added into AI algorithms.

TXE

^Well, that's just great.  

Perhaps it should be the case that AI should not be developed further without Artificial Morality (AM). I find the idea of an AI with no concept of right and wrong absolutely terrifying.

The article sounds very exagerated.

"Your microwave oven can learn with an added AI to suit to your lifestyle to decide when to go/off, therefore adquiring domination over humans."

Artificial inteligence needs more than knowledge to be one step ahead of us. Making self aware robots isn't enough. Humans becoming lazier as robots automatices our tasks will be a greater problem before losing control over AI.