The law seems to provide the perfect testbed to demonstrate the power and potential of expert systems and fuzzy logic to the mainstream software development community and to users. The law has structure, much of it in the form of a fixed framework established by statute, and much in the form of common practice which has evolved over the years. Unfortunately, in common law countries at least, an individual with every intention of complying with the law has a hard time looking up "the law". In addition to the access problem to statutes and administrative regulations, these official written expressions of "the law" are tempered and qualified by reference to practice and experience in the application of that statute to particular factual circumstances. (In fact, the law is an attempt by legislators to articulate the knowledge base modelling human behavior and the consequences thereof. The reason there is so much litigation and argument about application of statutes to particular
facts is that the representation attempted by the legislators is woefully inadequate. To be fair, the world is complex.)
Some areas of the law, such as corporate practice, rely on an acknowledged (although unarticulated) framework established by a combination of law and custom, with a relatively small margin of flexiblility for negoiation.
What makes a good lawyer? First, he or she must know the law (as promulgated AND practiced). Second, the lawyer should posses analytical ability, judgment, tact and negotiation skills. There doesn't appear to be any good reason for knowledge of the law to be a big mystery, solved only after years of practice. It should be possible (for a lawyer, if not a client) to look up the law, even if "the law" is not written in stone, but is probabilistic. It is the second set of characteristics which should distinguish good and poor legal practioners. Do we think any less of the physicist who uses a calculator, or the writer who uses a spell checker or grammar checker? What about the buyer for a retail store using an inventory control system? Financial analysts using computer-based financial models? Why do we insist that in certain disciplines, practioners must store the entire body of knowledge in their heads?
Because the law is expressed in natural language and not in the well-defined terms of science, it is much more diffucult to articulate the formulae and theories which exist in the law. However, this is also what makes the law such an exciting prospect for the application of AI techniques. Once expert systems master the hump of expressing and understanding natural language statements, their appeal to users from all walks of life should skyrocket.
The application of AI to the law requires the development of knowledge acquisition appropriate for legal practicioner/experts and knowledge representation schema appropriate to imprecisely-defined legal concepts and vague "rules" of law. The success of any such system also requires an interface layer customized to mesh with the experience, outlook, and training of legal practiioners. Such systems can also act as training devices.
Examples of application of AI to the law.