A Model for Ethical Artificial Intelligence

Abstract: The birth of the field of Artificial Intelligence was primarily due to the curiosity of researchers and inventors to create a perfect replica of the human mind. The thought of being able to simulate the human brain on a machine captured the imaginations of many, enticing and assimilating a wide array of research in the field, leading to a constant string of breakthroughs and new discoveries. With each breakthrough, our understanding of the boundaries and limits of the field were pushed, making it impossible to accurately estimate the complete scope of any project, with the only limitation being the technology available at the time. The primary purpose of any device implementing AI today, i.e. automation, was merely an application of the technology stumbled upon. As automation takes over our daily lives, it leads us to the main concern of where the line of ethics needs to be drawn and when does our pursuit for an ideal rational agent end with us questioning the morality of the decisions made by the machine. This paper tries to understand a model that weighs each choice with it�s moral conformation to truly decipher if a machine can understand each choice beyond the simple weighted values of desirability and goal reachability. After all, are we still trying to create the perfect human intelligence or just a perfect machine?

Keywords: Artificial Intelligence, Neural Networks, Brain Hierarchy, Morality, Ethics, Kant�s Laws, Responsibility.