"Could be dangerous?". Logic is something that doesn't care about anything but what logic says. Robots are not dangerous unless they're programmed by evil people to perform such tasks. The most important question is what is the definition of evil. What is evil to one person might be the business of other. Imagine many evil people having such so called AI bots.

Now, let's talk about the capacity and depth of this matter, suppose an AI bot codes 24X7 that means it can calculate all that goes around on this earth and beyond with in a week or so, on top of it all give it the power to code it's own logic. Let me tell you one thing you might be reading just the words when I'm typing this but there is no way on Earth anybody (or everybody ) can stop such madness, you can't fucking contain something of such massive scale. Man should stop messing around when it comes to certain things for the love of God.