Can a Disordered Mind by a Healthy Thing?

Can machines ever think and should we worry about a future where it may come to pass? Can the video games and computer programs that are marketed to be possessed of advanced artificial intelligence make moral decisions?

It was approximately 17 years ago that I had read an article by Marvin Minsky of MIT's Artificial Intelligence lab in which he pondered the ability of a computer or computer software to solve moral problems. The article was possessed of a disorder in composition that I believe betrayed the fact that MIT's A.I. lab wasn't making much headway in their quest to create a system that mimicked or actually possessed consciousness or could create solutions to moral problems. I wrote a letter to Mr Minsky detailing my perception of the problem. I believed that the human mind did work in a mathematical way that could be, to a degree, duplicated by a computer. The problem with computers duplicating higher planes of thought and reasoning like moral dilemmas was mathematical, it was however a mathematical problem that software or hardware could not tackle at that time or may ever be able to solve. AI is a featured component of video games these days, but they dynamically alter the processing of code and appear to solve basic problems not make moral decisions. On the contrary many of the video games rated for teen or mature consumers appear to be statically amoral.

When a human first considers a moral question and naturally takes a position based on learned morality or a long-standing moral precept, I believe that somewhere in their brain there exists something that could be recorded and measured mathematically and may one day be duplicated by an artificial system. The problem arises when the human considers other factors or even other moral positions that may be in conflict with the original problem or precept. When they join the two or more conflicting problems, mentally they corrupt the mathematics. The alteration of the value of a single number in a number system corrupts the entire number system. Regardless of how many equations created to allow for the alteration, the system is infinitely corrupted and infinite alterations would be needed to correct it. I believe the human brain experiences this same corruption. The original moral problem the human brain confronts could be represented mathematically. The conflicting moral conditions are considered and whoosh a number is substituted, the brain reels a moment, a person often senses the uncomfortable switch. It is not always as simple a something they had not considered; the mathematics of that part of the brains reasoning has been altered. The difference between the human brain and the computer is that the human brain, in most people, handles this very well. Considering the relative moral sophistication of our society and the scant protection of law in rush hour traffic, it probably does this hundreds of times a day.

Mr Minsky wanted to find an algorithm that could address moral problems. The algorithm may have existed when the human first considered the moral problem and an algorithm might then exist to produce a workable solution to the moral problem when that solution is at last considered. But the number systems, so to speak, that they are based on are not compatible. Intermediary calculations that would allow for this corruption or substitution in value may be infinitely complex.

As I have speculated above the human brain handles this very well, especially with moral problems. When it doesn't it produces some of the most abhorrent of mental conditions. A mentally retarded person can often handle moral problems, even if they approach the problem as a child would, there is no major corruption of the moral thought process. When highly intelligent people deal with moral problems or dilemmas ineffectively it could destroy or corrupt the person's morality. They function, sometimes at very high levels, due to the vast amount of stored knowledge and experience including acceptable moral behavior. But that functioning exists only to serve their corrupted moral imperatives. The mathematics of their higher thinking is dynamically corrupt, they can't handle the substitutions on the fly.

In one way of thinking a normal human brain considering a moral problem is disordered, at least from a mathematical perspective. Disordered, but still productively dynamic. An abnormal human brain considering a moral problem could be looked at as static producing a predictable result mathematically but not necessarily a moral one. This speculation has scientific backing on only one point, the inability of the scientific community to create a computer program or machine that satisfactorily provides dynamic answers to moral problems rather than canned answers from a database.

A machine making moral decisions does makes good science fiction though. From the HAL 9000 to the Terminator we are made wary of the morality of thinking or self-aware machines. I believe however this may play out to be a reasonable fear outside the realm of science fiction.