How computers can help draw the perfect contract

Contracts are becoming part of our daily life. While, until a few years ago, we signed contracts and agreements only when interacting with real life services such as with our bank, or the telephone company, today contracts pop up on our screen every time we install a new application on our computer or register to an online website or service.

Initial experiments have given promising results- Gordon Pace

If one were to dig beneath the surface of many such applications and online services one would find even more contracts – regulating what the online bookshop you use may and may not do when performing the hidden background transactions behind every order, with a bank to process payment, with third-party sellers, courier companies, publishers and so on.

The complexity of the interaction between such contracts is becoming difficult to manage. Yet, we somehow have to manage and process them, as failing to do so may have dire legal consequences.

Within the Department of Computer Science at the University of Malta, we are working on the use of tools to analyse the consequences of contracts and discover potential ambiguities within them. Although the research is still at an early stage, we have developed techniques which can be used to identify possible conflicts in contracts. These techniques have important practical consequences.

For instance, consider a computer system which uses different online bank services to provide users with a one-stop shop, and which chooses banks based on a user’s geographic location, current exchange rates, charges that bank levies, etc.

However, an additional complexity is that each bank service comes with a different contract or agreement. If a bank contract has a clause which requires that transfers above a certain value are made through a third party in a different country and another bank’s contract prohibits transfers to and from that country, it will not be legally possible to use the two banks to process a transaction. As these contracts grow, even lawyers may have difficulty identifying such conflicts, and all the while we are trying to see these processes automated.

The long-term aim of our research, however, goes beyond electronic contracts. In the same way that a CAD system helps a structural engineer to plan a building, calculate its structural strength and simulate its behaviour in different environments, and aids customers by providing 3D visualisations of the planned construction, we envisage computer-aided contract drafting tools which will play a similar role in the domain of legal agreements.

Such tools will support lawyers by providing automated analysis of the drafted contracts for conflicts, loopholes and ambiguities, and will help layman signatories by providing question-answering capabilities. Queries such as “Are there circumstances under which the seller may send my credit card information to another party without breaking this contract?” or “If I were to pay for the service twice, is the service provider obliged to notify me?” are answered.

Much of this is still a far-off dream, but initial experiments with extending logics to deal with such types of analysis, and the use of natural language techniques to enable automated processing of agreements have given promising results.

It is just a question of time until such tools will be ubiquitous among lawyers and notaries, just as spell-checkers and grammar analysis tools are among those writing documents today.

www.ridt.org.mt

Gordon Pace is head of department of computer science at the Faculty of Information and Communication Technology at the University of Malta.