Cognition vs Artificial Intelligence
Computer scientists love the idea of artificial intelligence (AI). It is the centerpiece of many mainstream science fiction works. It’s also a preferred buzzword of lazy vendors and marketers. Until computers can convince (trick) a reasonable human being that they are living beings (Turing test) all claims of AI are misleading at best. In this installment, I won’t debunk the types of claims of AI. We will examine the difference between how computers and humans think and the implications of the differences.
Digital vs Analog
Growing up in the Digital Age, I have been repeatedly taught that digital is better than analog. The stream of that marketing has been so deeply pushed on the populous that most people have a negative reaction to the word analog. It means old, inferior and lower quality.
The truth is, humans don’t normally operate digitally. Our world is filled with analog data. Colors are varying analog wavelengths. Sounds is the same with variance in modulation. Nearly every piece of data humans interact with is analog. If you put a digital signal (square wave) into a speaker it will make a strange sound before it destroys the speaker (melting the coil.) Our computer monitors have to output analog signals for us to understand them. Keyboards have to translate my analog input to digital so I can transmit this article.
While humans can understand digital Boolean logic (think flowcharts), computers cannot innately understand analog data. Input/Output (IO) devices convert the analog data that humans need to digital that the computer can process. Much research in the field of AI maps to solving the deficiency in analog processing for computers. Computers are very good at making mathematical and Boolean calculations. It works very well in absolutes but the human condition is rarely absolute. In an analog human world, there is a lot of variability between 0 and 1 (absolutes.)
In an investigation, a detective or a jury is rarely without any doubt. Business decisions cannot be made with dogmatic certainty. Art & music cannot be pragmatically judged as “good.” The human condition lives between the digital bounds of certainty. The human brain learns this before we can speak. Computers require a lot of work to emulate the behavior.
In the process of making decisions, there are different variables to be calculated. The number of variables and the options available to those variables in a calculation are known as dimensions. One of the primary attributes used to understand the complexity of a problem is its dimensionality. The more dimensions that exist, the more possible outcomes exist.
Making decisions on problem with limited dimensionality like “auto bill pay” is simple. There is limited variance in the amount of the bill. The bank account being debited doesn’t move/change often. On the other end of the spectrum is passing judgement on a murder trial. The list of variables include psychology, geography, politics, religion/philosophy, interpersonal relationship, finance, firearms, transport (and on and on.) Not only are the list of variables long, the list of possible options for each is long.
Risk of Error
Another consideration in automatic decision making is the penalty for being wrong. When IDS became IPS, business networks would routinely be crippled by valid traffic being blocked because the judgment logic (that ignored important dimensions related to decision making) was wrong. The same happened with antivirus when heuristic detection was implemented.
Role of Each
The computer is extremely powerful in its ability to calculate information very quickly. The human brain is in a constant state of learning that can combine information from different domains. The human brain is not as quick as a computer but is much more resilient, agile and innovative. This is one reason that cyber-criminals have been so successful over the last decade. They recognize the superiority of the human mind to adapt faster than computers. If military OODA loops apply to cyber-crime, a human brain that is assisted by machines will always out position machine only adversaries.