Friday, March 20, 2009

Connection between artificial intelligence and law

According to Patrick Henry Winson, "Artificial intelligence is the study of the computations that make it possible to perceive, reason and act." [Artificial Intelligence, Winston, p.3]

Now, what is perception?

The best answer to that is found in Tarka Samgraha by Annambatta (translated by Swami Virupakshananda):

"Perception is the apprehension born of the contact of an organ with an object."

Further, sense organs are the seats of perception and intelligence. To “sense” is to “know”, have “knowledge”. Sense of touch, taste, smell, vision, hearing, and reasoning are commonly attributed to humans. Machines could have more accurate sensors to detect other wavelengths and physical properties.

What is apprehension then? It is explained by giving its two kinds: Indeterminate apprehension: knowledge without any attribute (eg. this is something) and Determinate apprehension: Attributive knowledge (eg. this is a laptop, a desktop computer, a mainframe).

To Reason, is to test validity of perceptions and make inferences from what is known with an object or purpose.

Action is through movement, speech and other means: animate or inanimate.

Attributive knowledge about a legal system is a sine qua non for artificial intelligence at the level of humans. Marvin Minsky, co-founder of one of the oldest AI Labs in the world at MIT, wrote why "AI is brain dead" at http://www.wired.com/wired/archive/11.08/view.html?pg=3

Wired: The biggest name in artificial intelligence declares AI research "brain-dead" since the 1970s. What gives?
Minsky: There is no computer that has common sense. We're only getting the kinds of things that are capable of making an airline reservation. No computer can look around a room and tell you about it. But the real topic of my talk was overpopulation.

What's overpopulation got to do with AI?
The elderly segment of the population is growing to the point where there won't be enough doctors, nurses, and nurses' aides. We should be working to get robots to pick up the slack.

What's AI's biggest deficiency right now?
The lack of people with an interest in commonsense reasoning for computers. I've found maybe a dozen. Douglas Lenat's Cyc ["psyche"] is the only major program that has collected commonsense knowledge. But it's not there yet.

If AI's brain-dead, aren't you partially to blame?
No. I've been continually working on the problem. I'm trying to put a new project together, but it's hard to get 10 capable people. It would take five or ten years, and nobody wants to put that kind of time in - people want to double their money overnight.


Let us assume a machine well equipped with vision, hearing, taste and other sophisticated sensors. How would it have any "common sense"? If it has a database of rules relating buildings, there is a chance it could start making comments on the architecture of the room. To begin with, it would be able to talk about compliance, and if it has studied enough building plans, it could tell us what it thinks, just as any of us.

The connection between ai and law have never been deeply appreciated. The comment on computers making "airline reservations" betrays that. I could safely say a database of legal rules, i.e. legislation, is in fact a repository of common sense rules, and porting law to computers is a basic step in making artificial intelligence even possible. To be fair, Minsky himself has said elsewhere "Societies need rules that make no sense for individuals. For example, it makes no difference whether a single car drives on the left or on the right. But it makes all the difference when there are many cars!" The rule set forming the background for human transactions cannot be inferred by machines on their own. They need to be ported.

The fair rent calculator script took three months of legal research. It took only a week to write the javascript. That was just one provision relating to rents in this part of India. The globalised world would need every provision as machine readable code. If the legal fraternity gets involved in porting law to computers, it would benefit the legal profession and also help solve the ai problem that is pending for the past several decades. If law is ported as machine readable code, ai would be a reality within a decade from now.