AI is not just a technical topic, but more of a science. The engineering of making intelligent machines, as the book says. The complexity of AI is extended to the brain level of understanding the human intelligence. “The computational part of the ability of achieving goals” is the current best used expression to define the word ‘intelligence’. Goals are the reasons of life. Thus, every living creature can be measured in terms of intelligence. While the term intelligence is still generally applied only to humans, it is believed that we can come to a definition that will identify it in terms that are not specific to humanity.

The first goal of AI research was to simulate human intelligence in agents known as robots. The failure to compile a successful project to demonstrate such theories is due to one problem; the comparison between human intelligence and AI lacks an understanding of the intellectual mechanisms that form the building blocks of what we know as intelligence. Simply put, we cannot create a totality out of something that we know very little about. Even worse, we don’t know how much we know in our supposedly complete image!

In order to achieve the goal of imitating human intelligence, researchers believe that new fundamental ideas are necessary, and therefore it cannot yet be predicted when human level intelligence in an agent will be achieved. The philosopher Hubert Dreyfus said that AI is impossible. The computer scientist Joseph Weizenbaum said that the idea is obscene, anti-human and immoral. Others have said that since artificial intelligence hasn’t reached human level by now, it must be impossible. But, is it?

The Three Laws were created to ensure sufficient control over robots, and prevent robots from harming humans. The Laws also make sure that robots must obey orders given, and lastly protect itself.

Law 1 : A robot may not injure a human being, or, through inaction, allow a human being to come to harm.
Law 2 : A robot must obey orders given it by human beings, except where such orders would conflict with the First Law.
Law 3 : A robot must protect its own existence as long as such protection does not conflict with the First or Second Law.

The failure of these laws is illustrated through the results of various situations. One being, the incapability of understanding how an intelligent robot would interpret the words: “harm” or “injure”. For example, if a human ‘wants’ to die, is killing them actually causing them harm or injury? This illustrates a fundamental problem between foolwing orders and preventing injury.

The other problem regarding interpretation is that these laws are only applicable to human beings. How about if a robot sees that your car is dangerous? Then, to avoid violating the Laws, the robot will disassemble your car to prevent an accident.

“No matter how one creates a moral center for an AI, that moral aspect will only exist so long as it remains advantageous for the AI to maintain it”. The laws fail because, simply, they are too simple. They only describe a minimum set of instructions to follow.

The laws also ignore the ability of the robots to reprogram themselves. If you are an Asimov fan, you would know that this happened in one of his novels (1). This is an inevitable consequence of an intelligent core. Don’t you often ‘change’ your principles or your views, considering that you understood that the opposite of that principle is bad? On what basis did you evaluate it being good or bad?

In that novel (1), Daneel created a higher order Law, called the Zeroth Law, which states that “A robot cannot cause harm to humanity, or, through inaction, allow humanity to come to harm”. With a small amount of thinking, you’d get that this law would allow a robot to destroy a human if they feel that human was causing harm to humanity. Furthermore, as mentioned above, the definition of “harm” is variable to a robot. Thus, the circle of evidence would be perfect.

The real problem with the simplicity of the Laws originates with the complexity of the human brain. The idea behind it is that human beings are not ‘just’ intelligent. We have other facets that make us what we are. Feelings, for example, are a critical part of the argument that no matter how intelligent a robot is, it is still impossible that the robot can feel.

“I guess to put it simply, it doesn’t matter what the text of the moral code is, because that can be modulated easily. What matters is the necessity behind the moral code for the well being of the AI”