Semper letteris mandate
Millions of years ago, out of a primordial sludge consisting of water, saline, and amino acids, the first life forms began their long struggle to evolve. Through random chance and natural selection, some of these primitive life forms became rational, thinking beings. Others became Howard Stern.
Today a similar process is taking place right here on the Web. Out of a primordial sludge consisting of chat rooms, tiny wireless cameras, and fan pages for the Spice Girls, a new life form is arising which may well evolve into a new kind of rational, thinking consciousness. Or get its own nation-wide radio show on which it says “tits” a lot.
The force behind this new brain is the Mindpixel Corpus, the brainchild of AI enthusiast Chris McKinstry. With the vaguely perplexing motto, “Making the implicit, explicit,” its aim is to gather billions of simple observations, true regardless of race, gender, or individual differences, which will form a working model of the human mind.
Visitors are encouraged to submit a “mindpixel.” A mindpixel, they explain, is “a binary statement of consensus fact such as ‘Water is wet’ or ‘It is difficult to swim with ski pants on’.” Upon submitting an entry, the visitor is then asked to rate ten previously submitted mindpixels according to their “truth” and “value.” The plan is to gradually provide the “brain” with countless units of human experience, graded according to consensus and reliability.
The brain being created in this grand experiment is called “Generic Artificial Consciousness,” or GAC (pronounced “Jack”).
The project head is Dr. Robert Epstein, whom the Mindpixel site calls “one of the world’s leading experts on human and machine behavior.” Dr. Epstein has a psychology doctorate from Harvard (1981), is the founder and Director Emeritus of the Cambridge Center for Behavioral Studies in Massachusetts, and is Adjunct Professor of Psychology at San Diego State University.
In its first year on-line, the Mindpixel Corpus has received nearly 8 million individual measurements of more than 355,000 individual items of “human consensus experience.” Upon completion of its collection phase in 2010, work will then begin to create a statistical model of an average human mind with the aim of using it as a foundation for true artificial consciousness.
All of which is possible only because of the World Wide Web. Without the aid of the Internet, the data entry alone would have cost $250 million.
Will it be successful? To his credit, Dr. Epstein admits he isn’t sure. “We don’t know if it is possible to build a normal personality out of millions of little pieces. This experiment will tell us how reasonable the idea is.”
Of course, whether or not it succeeds depends largely upon how we define “success.” If by “success” we mean the creation of a truly artificial consciousness, many experts believed it is probably doomed to failure. They claim that electronically storing millions of simple statements about experience, and processing them according to an established set of rules is no more likely to produce consciousness than doing the same thing with statements written on pieces of paper.
One long-standing argument against this approach to Artificial Intelligence is John Searle’s “Chinese Room Argument.”
In this thought-experiment, a man is locked into a room with nothing more than a book of complex rules. Through a slot in the door come slips of paper upon which are written Chinese words. The man compares these symbols with his rule book, writes down the result (which is an English translation), and shoves it back through the slot.
Searle’s point is that even when you add a living human being to it, no rule-based system is conscious.
But even if the GAC fails to reach a state of true consciousness, it still holds out hope of offering invaluable insights into the way minds work.
If nothing else, GAC could completely revamp the way we conduct polls and surveys. No need for dozens of employees to phone thousands of average citizens to discover which product name they like better: we could just ask GAC.
Of course, we’d have to know exactly what GAC’s “demographic” is. In other words, we’d have to find out just what kind of “person” is residing in GAC’s virtual mind.
And this is precisely the purpose behind the months-long psychological test GAC is presently undergoing. When finished, not only will we have a window on its underlying personality, but GAC will be the first machine-based artificial personality to be tested by the MMPI (Minnesota Multiphasic Personality Inventory), used in both corporate hiring practices, and criminal court proceedings.
It will be interesting to see whether GAC turns out to be an ideal CEO, or found “not guilty by reasons of insanity.”
[Post Publication Note: While on the surface, Mindpixel appeared to be a positive and even light-hearted project, its underbelly was far darker. While Dr. Epstein was the project leader, the actual concept of Mindpixel was Chris McKinstry, a troubled but brilliant AI specialist. The Mindpixel project floundered and was put on mothballs only a few months after this article was written, and a few years later, McKinstry committed suicide (an act he had attempted several years earlier in a well-publicized incident in Toronto). Very shortly after this, another of AI’s most prominent promoters and rival of McKinstry, Push Singh, also committed suicide.
You can read the full story here: Two AI Pioneers. Two Bizarre Suicides. What Really Happened?