Drafting of legal and quasi-legal documents such as commercial contracts, insurance policies, license agreements, employee (or student) handbooks, benefits summaries, and consent agreements is a resource-intensive process for many companies and government agencies. The proposed research applies artificial intelligence problem solving, natural language processing, and machine learning techniques to the problem of automated document drafting.
We are developing a model of document drafting as a problem solving activity in which the following knowledge bases are integrated:
This model provides an architectural framework for an intelligent drafting assistant system, and we are developing the algorithms required to implement the model in a working system.
A drawback of many artificial intelligence applications is the effort and complexity of developing and maintaining the knowledge bases. Therefore, we plan to explore the potential for applying natural language processing and machine learning techniques to automate some of this effort.
Although there is potential for reducing the cost of document drafting as a result of automation, the greatest potential benefits to users are the reduction of errors (thus avoiding costly negotiation or litigation), and much shorter turnaround time for preparing or modifying documents. For routine transactions such as real-estate closings, drafting automation can reduce the time required from a day to a few minutes, and for complex transactions the potential reduction in turnaround time can be even more dramatic.