Why Artificial Intelligence is a myth.

 

Computers will never be able to think and reason, they are fundamentally rule based, and computational machines. While they do this extremely well, better than any human possibly could, they are fundamentally limited (and always will be) to the rules that they are programmed with.

 

People and organisms that can think, do not operate on rules, but on functionality or purpose. This fundamental difference is a hurdle that can never be over come. The world is infinity complex, and no amount of code or rules will allow a computer to overcome and adapt to the endless possibilities. Thinking organisms have evolved to be able to filter out information that is not useful to them in real time and instantaneously. What is more is that this filter is constantly changing depending on the real time needs and wants of the organism.

 

Let’s look at a very basic example to illustrate this, something as simple as a rock. It will be overlooked constantly and have no significance at all hence filtered out, now lets say your bored you and your friend decide to start skipping them across a lake, it is now not a rock but a toy, next you need to open a shell to get a nut, now its a tool or hammer. If a bear suddenly comes out of the woods, its a weapon, and after a long day it becomes a chair. So something as simple and as mundane as a rock really has countless possibilities of fulfilling different functions, and it is not the rock that changes but our needs. This is true about everything. It is just not possible for something that is rule based and computational in nature to operate in a functional/purpose driven state (except in a very limited fashion and at the expense of many teams of people spending countless ours to design them to preform a small set of functions). There are simply too many possible combinations of needs and solutions to possibly ever be put into a set of rules that a machine could use to think.

 

Things like the Turing test (a computer can talk to a human and the human does not know it is a computer but thinks its another human) will never be achieve, they can make as large of a database as they want but it will never be able to filter out all the subtle non verbal or even connotations, context, and slang in written language. Communication alone is far to complex for a rule based machine to understand. The rules and language evolve so fast and vary so much from group to group. So a large enough database might fool some humans into thinking they are talking to a human some of the time, on a very limited conversation. It will never be able to actually understand what is being said and able to adapt to the changes in language or time and from group to group.

 

So while the media loves to make a big deal out of a computer beating a human at chess or go, it is nothing more than brute force, based on a set of rules. In chess this is perfectly clear that it was brute force. In the Go case it is still brute force just by different means, it used a large database of past games played instead of raw computational power. Neither of which are impressive or demonstrate thinking.

 

What is far more frightening than “thinking” computers, is dumb ones with unethical government/corporations/organizations programming them with very simple tasks along with mass production. It would be very easy with current technology to build and program a very dumb machine to recognize a person fly, roll, drift, walk, etc. and say set of a small charge of high explosive and kill them. Now if you mass produced 100’s of millions of these you could easily do so for a few billion dollars with current production technology. I have a bad feeling that this will eventually be the way that warfare will go, low tech/cost drones of death.

 

Artificial Intelligence is a distraction from the real threat that dumb low tech killing machines actually pose. I am not an expert in A.I. if I am overlooking something please correct me on it as I love to learn new things.

 

Thanks for reading :)

 

Sanity Unhinged