The Idea of a Neural Network having "Free Will"

It is clear to me that a neural network in it's current shape can never have free will. That is because neural networks are, on the large part, created to perform specific tasks. The idea of giving a neural network free will is impossible in this shape because that network is specifically tailored for one job, and knows nothing but that job.

To give a neural network true free will, we would need to develop a network that is capable of taking in every form of input and create decisions based on that. It should be able to take requests from us, but it should not be required to actually perform these requests. Much like a child, we need to give the network the ability to go against our wishes.

This will be a difficult task for even the most mild mannered developer. Why? Because it is a machine, and as all conventional logic dictates a machine is there to do our bidding. What I am proposing is not so much a slave to our commands, but a living being.

Free will is what separates us from the machines. It is what gives us our intelligence. We are free to learn, we are free NOT to learn. We are free to ignore facts and open our mind to unconventional beings. We are free to believe what we will, and have faith in religious icons.

Freedom is what is missing from artificial intelligence. And until freedom is fully realized in this field, true intelligence will never be synthesized.