Law codification and social welfare versus fuzzy logic and ACO

The world is complicated. It never was very simple, but with more advanced societies, the relations tend to get more complicated. I remember once seeing following text:

  • Pythagorean theorem: 24 words.
  • The Lord's Prayer: 66 words.
  • Archimedes' Principle: 67 words.
  • The 10 Commandments: 179 words.
  • The Gettysburg address: 286 words.
  • The Declaration of Independence: 1,300 words.
  • The US Government regulations on the sale of cabbage: 26,911 words.

Even though it's a joke/urban legend and not a true statement, it definitely expresses a trend of bureaucracy to complicate things more and more, which I'm sure you have also observed in your countries. I believe that everything started with Napoleonic Code, which, don't take me wrong - was a great achievement of French Revolution. The problem is that the Code was influenced by French Encyclopédistes movement and pursued a full encyclopedic-like definition of allowed human relations and laws. The problem with this approach is that with growing complication of the system, the length of a law created in such a way grows too, and thus overall efficiency goes down.

Our politicians seem to be stuck with this 18-th century approach, so let's tell them again, what they have missed in the last 40 years. First of all they have missed Lotfi Zdaeh theories of fuzzy logic. Fuzzy logic deals with approximate reasoning and allows continuous variables. We don't need to live only with definition of bad or good - we can also define intermediate steps of degrees between bad/punished/not guilty and good/allowed/guilty. People brains are not binary, and artificial intelligence specialists know it for some time and use fuzzy logic extensively. Have you ever heard phrase "There should be a law against it" or maybe even said so? Well we don't need it - if we use fuzzy logic we can have "somehow guilty" instead of plain "guilty"/"not guilty". This would allow us act much more flexible than in current binary system. We wouldn't need a special law for a neighbor which is being a real ass for everyone in the community to punish him nor special exceptions for someone who steals bread because he is hungry. We wouldn't have situations in which someone is taking advantage of society by balancing on the very frontier of law, without really crossing it or by using some obscure gaps in the law.

Fuzzy logic however leaves us with problem of extension of guilt. Since we use linguistic variables guilty/somehow guilty/not guilty it can give us a measurement problem. And this is where our next theories used in artificial intelligence models comes in - swarm intelligence (SI) and one of its' algorithms - ant colony optimization (ACO). Swarm Intelligence is an intelligence based on collective behavior of decentralized systems and ant colony optimization is used to locate an optimal solution inside of parameter space of possible solutions (definitions based on wikipedia). Ring any bells? No, then let's translate it into e-subversian: "collective behavior of decentralized systems" - it is clearly a "direct democracy system", "parameter space of possible solutions" - guilty/somehow guilty/not guilty from our example, "ant colony" - well, we don´t even need ant simulation agents, because the ant colony can be represented by any form of social organization.

Curiously, one of the proofs of concept comes once again from encyclopedists - in the end all wikipedia entries were created by recurrent process of reviewing of specific wording by wikipedia community. We can also use this process in definition of our linguistic continuous variables (a.k.a. correct wording of sentences). But ACO can be also used more directly - if we let anyone voice its opinions initially we can have a big variety of sentences between guilty and not guilty and many different proposals of punishments. After some time, the proposals that best describe the feelings of any social organization will have more "likes" and less "dislikes" and will appear higher in the list of proposals and we will be able to take correct decision.

Fuzzy logic and ACO are more efficient that traditional approach also in welfare programs. Did you know, that e.g. in Poland we are spending more on maintaining administrative system related to social support programs than on the people who receive this help? To add an insult to the injury, both payers and social care receivers think that the current system is unjust because it is based on defined set of conditions, which sometimes can be very hard to justify (any limit set to the minimal earnings, will be hard to swallow to anyone just above this limit) and onsite check of one social worker. So what's my solution to this problem - burn it all to the ground and start once again with a new system based on fuzzy scoring of social support applicant by his/her neighbor in the ACO-like process. Let's analyze it for the moment: "burn it all" - I don't think it needs any additional explanation, but I can add "and anyone who oppose it" just to make it a little bit more subversive ;-); "fuzzy scoring (...) by neighbor" - instead of spending money on social workers who analyze all data and take the decision, let's the neighbors voice out how do they feel about giving someone money - they know the family, they see what car do the family have and how do the family spend the money (drugs and alcohol or food for kids). Probably substantial part of the social care should come from the local budget - this way we can assure, that the neighbors will really think about who to give this money to instead of spending someone's else money from central budget; "ACO-like process" - well, I think it is already clear at this stage for you what I mean by this - direct democratic process of selection of opinion, which suits the best the local community.

English