Why can't we develop Artificial Intelligence yet?

• Muhamad Rawi
• 09/09/2018
• 4485
• 5,394

In 2017, the artificial intelligence department at Facebook has performed some weird experiment where two intelligent chatbots were put to negotiations about some deal, at the beginning the two chatbots were speaking normally and all of a sudden both of them started using some weird language with repetition of certain meaningless phrases. Facebook scientists couldn't interpret exactly what happened and they decided to shut off the experiment, the conclusion was that the two chatbots have developed some sort of shorthand language to help them communicate better. We don't know exactly what happened but we don't lean towards believing  that this was the case anyway, after all the experiment has failed. Here is a link of what happened during this experiment in this article with the text of full conversation.

Meanwhile, the robot (Sophia) was running around the world making some seemingly smart conversations on public televisions. Sophia pretends to be a smart AI and it made a whole lot of buzz among common people who seemed to finally believe in the probability of artificial intelligence. The Head of AI department in facebook (Yann LeCun) stated that Sophia is probably a hoax, Facebook's Yann LeCun said Sophia was "complete bullsh*t" and criticised media for giving coverage to "Potemkin AI. Some other critics claimed that Sophia is like when some big corporations are trying to invent a car where they stand at the tire development phase and then all of a sudden someone claims he has invented a spaceship!. Hansen Robotics (the head company for Sophia) somehow keep making interviews all over the world through Sophia that successfully managed to acquire the first robot citizenship  in history (Saudi Arabian citizenship). We have previously discussed the Sophia case in this article with more details.

Great companies are working on AI development for ages now, IBM, Facebook, Google, and many other big companies and government corporations. They have all failed till now or at least they may have achieved some limited success in certain fields but could any of these big corporations claim that they have developed a complete humanoid AI? the answer is certainly NO. In this article we will try to discuss the impossibility of developing real humanoid AI and why all these companies have failed.

What is Artificial Intelligence?

Artificial Intelligence (AI) is a broad field of science specialties that care for developing smart systems and machines capable of logical thinking and decision making based on proper logic and available data. According to techopedia:

 Artificial intelligence (AI) is an area of computer science that emphasizes the creation of intelligent machines that work and react like humans. Some of the activities computers with artificial intelligence are designed for include:
1- Speech recognition
2- Learning
3- Planning
4- Problem solving

Now we have to ask this important question: If all these previous skills are available these days, Why can't we develop AI yet?

You might feel familiar with the famous google translate bot and personal assistants like SIRI but all these bots seem to be a bit stupid till now, if your primary language isn't English (like me) and you try to translate a whole phrase using google's translation bot you will easily fall a victim of google's mistranslated phrases. SIRI's ability is still limited though and it can hardly answer direct questions only through searching the internet, it can hardly understand what you are saying and will give you mostly wrong answers. Android's personal assistant is even far behind SIRI and you can rarely find anyone using it, I even forgot it's name due to it's massive unpopularity.

In the past century, (Alan Touring) the father of modern computing has given us a simple but strong definition of AI through his famous (Touring Test). You can identify true AI through chatting with a bot not knowing it's true identity and when you can't identify that the other side is a bot not a real human, that's when we can claim true AI. However, Touring test could be cheated nowadays like what Sophia did to the whole world but not to AI experts who can uncover the true identity of a robot but asking specific questions that only can be answered by humans.

What is human intelligence?

We can generally define human intelligence as the opposite of intelligence itself in general, human intelligence depends on many components of human psychological nature in addition to the mind, the consciousness, and finally the only physical part of human intelligence which is brain and neurons. We can't state by any means that human brain in it's physical form is the only factor determining the state of intelligence, there is also the psychological component as human intelligence's main drive for all human acts and tendencies. Humans are intelligent because they are incomplete and they are always seeking perfection, there is always something missing in each of our souls and minds and we are always looking towards the missing rock in our perfect wall. Some people seek perfection for this life and for those who believe in a second life (like myself) we seek perfection in our eternal life, everyone of us is looking for something and hence we have the drive to act intelligent.

Humans are intelligent with their consciousness, their incompleteness, and hence their life goals. Humans differ from animals in their need to get more of everything, their ability of virtual simulations for situations that would happen in certain cases, and their inner ability of learning and acquiring new skills and experiences. However, the major component of human intelligence is (awareness), humans possess awareness of surrounding objects and other humans, they are aware of themselves and the whole universe around.

Humans are also prepared to make brave decisions that are not based on crude logic like self sacrifice for the sake of their home land. Humans are capable of loving and hating, humans need to feel safety and company, humans need to breed their offspring for the sake of immortality and continuity of their names. Humans possess the full freedom of taking or leaving, stealing or working, justice and injustice, or even believe or disbelieve. How could a computer believe or disbelieve ?!.

Humans are even capable of not giving a decision at all, you can give a human the decision to select between two objects and they would even prefer not to make that decision at all, humans have the ability not to make the call or stay at the neutral zone without a specific decision. Humans can think out of the box with no limits to their creations. We can perform a non logical experiment or even expose ourselves to danger for the sake of higher psychological purposes. Even the most materialistic personalities could transcend their body needs for a higher purpose, humans do that all the time.

We didn't exaggerate when we claimed that true human intelligence is the opposite of intelligence on a general prospective as humans can make non logical decisions all the time that no AI can make and that would lead to their dissolution eventually, robots don't smoke or dink alcohol for example but humans do that for the sake of psychological satisfaction and mood improvement. Humans could travel to another company for the sake of pleasure, why would a robot decide to spend money for pleasure?. Humans could spend a lot of money for the sake of making someone else happy (love - marriage), why would robots spend their money for a mostly losing deal that won't make him any benefit unlike humans who would wake up early and go to work for the sake of making money that would help them raise their kids. Human life is full of non logical decisions that would be very hard for a robot to understand fully.

Humans are just a physical form or more than that?

The materialistic western mind deals with humans as a solid physical form, you can't find any claims for soul in any scientific paper. We understand that human soul cannot be seen or tested or experimented at any scientific measures and I quote this verse from the holy Quo-ran (and they ask you about the soul, say to them: the soul is only a matter of my god and you were not given but few knowledge). Even religions claim that the souls is non materialistic and cannot be seen or examined but does it mean it doesn't exist?. Modern human materialistic mind has chosen to ignore the soul completely and decided to deal with human beings as just meat and flesh (Brain, neurons, neurotransmitters,  and hormones) with nothing more.

For the sake of modern materialists minds to properly define human intelligence or the whole human psychological process many trends tended to interpret all psychological disorders with organic causes. Now let's assume that there is some guy who is obsessed with murder (serial killer, rapist, or even a mad military commander who finds it easy to kill civilians), a materialistic mind would want to interpret this guy's sickness with organic causes but can we really do that? do you know any organic differences in serial killers brains than yours? could you provide any materialistic evidence of this claim that mad serial killers are different in their brains from ours?. The answer is and will always be a big NO, you can even measure all related psychological behaviors either there were accepted or not on community level (killing, rape, sexual harassment, and tendency to hurt others) they all belong to the same family of bad human practice and they would all look extremely normal with brain CT or MRI.

The materialistic theory have failed completely to prove any proven or experimental evidences of organic drives for these sick behaviors, do we need more than brain cells and neurotransmitters to fully explain human behaviors? I think we do.

Human consciousness:

You can read this article here for a clever trial to define human consciousness with some useful information about the complexity of human consciousness and how hard it is to fully understand and even define it. Many scientists now are suggesting (Dualism) as a better alternative than just the physical definition of human nature. Dualism simply means that humans are both a physical material + something else, you can call it soul, energy, or whatever terms you would like to use but it is there anyway. Even in physics (Quantum Mechanics for example), it is now a fact that micro physical particles have both physical and non physical forms. Electrons and photons for example could be both matters and energy at the same time, read this article we wrote about quantum mechanics and how did this science change the way we look at our existence.

The truth is: Materialistic minds have lost a lot with eliminating non physical components of human psychology and we think that for this particular reason; all attempts to develop a true artificial intelligence have failed.

What is the difference between humans and machines?

There are many differences between human minds and machines considering the prospective of intelligence and decision making steps. for example:

1- Machines can't act just by themselves, they will always need an external motive to perform some action. Pushing the power button on your laptop or cell phone pushes your machine to power on and start operating, that was an order at the simplest form. You will also need to give your computer some orders to make them do some math calculations or open the internet browser, even the whole artificial intelligence utilities in social networks or search engines or even that camera that pictures you while crossing the red light with your car, all these systems will act upon some order or even they watch for a specific change at the regular pattern. These systems can't act by themselves and they will need to be pushed to work by their creators or even they will be programmed to perform some actions based on your actions, you do the action and they will react but they will never act on their own even if they seem to be.

2- Machines can't think outside the box and they will only make decisions based on some range of options they can pick from. The range of machine actions must be pre-programmed by human developers and so they can't make or create new decisions on their own. So based on the ability of the programmer to give a wider range of options to the computer system they will act on wider scale but nothing outside that closed box of actions.

3- Humans can misbehave, they can cheat or lie or do any kind of misbehavior. Computers can't do that simply and although you would hate these previous actions and all kinds of misbehavior generally but they are necessary for human life. These characters are strictly human and they help humans all the time to bypass awkward situations, survive fatal conditions, and even make profit using sneaky ways. Humans can compare in their minds to behave appropriately at the moment or misbehave for a better gain in the future, (end justifies the mean) didn't we all say that?.

4- Humans can answer with yes, no, maybe, or even any other set of answers that don't conclude a strict or definite decision upon some calls but computers can only answer with yes or no. Computers are not built to make decisions that are not based on a complete knowledge or awareness of all conditions, even computer randomization and generating random codes or numbers are not truly random, you will know that fact if you specialize in computer programming field. Computers can act upon a specific probability (15% or 70% for example) but they can't make calls for probabilities less than 50% based on pure belief. Computers can't gamble or make stupid calls with weak cards unlike humans who choose to gamble and make the same stupid calls a computer wouldn't do for the sake of a big prize.

So, can we really develop AI one day?

AI is already there and it is around us everywhere. It exists in your cell phone, your computer system, your social network account, your youtube experience, and in almost every action you make on the internet nowadays. They collect information about your preferences, your search queries, places you have visited and the ones you would like to visit, your preferred drink, you religious believes, and even your political views. AI tries to answer your questions and fulfill your needs, it even tries to catch you with new needs you don't know about. However, all these actions are pre-programmed for a certain purpose and can only be driven by their developers, these systems don't have a personal entity and they are not truly humanoid although they pretend to be.

The truth is that developing a true humanoid AI requires catching with human nature with all advantages and disadvantages. To develop a true AI you will need to make it able to negotiate, cheat, lie, forget, forgive, seeking self achievement, and also self denial at specific situations for the sake of a higher purpose.

Humanoid AIs should acquire moral drives and state internal regulations among similar entities (AI communities). A true AI would not be isolated from a herd or community of similar objects of the same type or different types, it will need to be a part of a group which will eventually lead to a hierarchical system of authority and power, it will need discipline and regulatory laws which should also contain coercion and punishment acts for those who disobey the law or decide to break it and it will happen if they possess real free will. These all facts should be fulfilled to develop a true humanoid AI and not just a smart chatbot that searches the internet to answer your questions. It is very complicated and hard to develop AI and we might not be able to develop AI soon enough but when we do the life as we know it will change completely with the unlimited abilities this new system will acquire.

Artificial intelligence and human intelligence don't meet, they might even be at the opposite of each other and when humans can give away their moral and non moral acts and decide to work as strict machines; this will be equal to the probability of developing a true artificial humanoid AI.