American Patriot. You are part way right. we are the one that write the code and the algorithms, but AI is far more advanced than a string telling it do specific commands. How many times have you tried to used a program, but it doesn't work correctly or not at all, and then you install some updates and it works fine after? codes need to be changed and upgraded when new information is available. Then there is also a theoretical AI process called rampancy, which is the uncontrolled expansion and self awareness of an AI. Also if there are robots or androids running around, then, (like almost all o f our computers) they are susceptible to malicious viruses/spyware and other vicious attacks. In short there are many things that could possibly lead to a sci-fi-esque uprising of AI drones, but odds are we will take every precaution for this not to happen.

Whether it's like HAL or Joshua, that is up for us to decide. Even if these machines do exhibit AI, they are still machines, and are subject to control by their own code (which we the humans create for the bots / androids / computer to use). Sci-fi flicks where the AI drone goes haywire and tries to kill all humans is just that sci-fi. The AI to serve us will not be made intelligent or be allowed to be sentient enough to recognize us as a threat, or even know that we are their masters. The AI of the future should be more like  an intelligent pet.

Congratulations - The "if - then" theorem when applied on "many levels" of understanding can bring about results not thought possible just a coujle of years ago.