A Deconstruction of the 3 Laws of Robotics

(Disclaimer: Couldn't get videos to work but links will open up in another tab.)
Isaac Asimov's 3 Laws of Robotics are bound by the etymological (study of the meaning of words) definitions of words found within them. Furthermore, different unanswered dilemmas or moral questions must be programmed into the robot, thus making it susceptible to human subjectivity. For these reasons, the 3 Laws of Robotics are more complex than they initially appear and should not be used as a means of controlling robotic behavior.

The 3 Laws of Robotics are stated as follows:

1. A robot may not injure a human being or, through inaction, allow a human being to come to harm.

2. A robot must obey orders given to it by human beings except where such orders would conflict with the First Law.

3. A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

If you take the meaning of 'harm' or 'injury' to not only be physical, but also psychological, then for a robot to do the utilitarian calculus in determining how not to injure a human being, it would need a strong grasp of the nuances of human psychology. More importantly, it may be true that no matter what the robot does, it will either injure a human being or allow it to come to harm through inaction.

Take the Trolley Problem for instance.

 In this case, Isaac Asimov's First Law of Robotics cannot be used in forming a decision. If the robot were to pull the lever, it would divert the trolley track and kill the single person, violating the, "A robot may not injure a human being," clause. However, by not pulling the lever and allowing 5 people to die, it violates the clause, "or, through inaction, allow a human being to come to harm."
While extreme scenarios like this one would not be a day-to-day occurrence (I hope), lesser scenarios would arise where unfortunately, this dilemma would still occur. Although, let's give the First Law of Robotics the benefit of the doubt and divide it semantically.

Let's take the First Law of Robotics and splice into an 'either-or' statement:
 1. Either, "A robot may not injure a human being, "
 2. Or, "through inaction, allow a human being to come to harm."

This gives a robot a choice, therefore dodging the Trolley Problem scenario. Yet, this creates a new problem: In order for a robot to make a decision in this case, it would need some way of evaluating which side of the 'either-or' statement it should take, inevitably leading to some human assistance. This is where the Second Law of Robotics comes into play; this 'evaluation process' could be an order given to it by a person through the Second Law of Robotics that rather than 'conflicting' with The First Law of Robotics, makes it possible instead. Moreover, the Third Law of Robotics can work because of this 'either-or' relationship. Nonetheless, this is flawed, but for an entirely different reason.

The robot needs an internal understanding of what 'harm', 'injury', or 'humanity' is. Etymologically, these are difficult to define and span the entire range of Ethics and Moral Theory.

This argument is best explained in this Computerphile video.

For these reasons, the 3 Laws of Robotics are more complicated than they seem, seemingly denying The Laws of Robotics' simplistic nature. Henceforth, they should not be used as a means of controlling robotic behavior.