This site was created by, and is maintained by, Selmer Bringsjord, Director of RPI's Minds & Machines Program. Copyright 1997, Selmer Bringsjord.
Overview || Project Leaders || Team Members || Specific Sub-Projects || Relevant Links || Contact Info
The senior associate at the prestigious Wall St. law firm looks at the desk clock and smiles. Four o'clock on Friday and a blissful work-free weekend awaits.... Just one more hour until freedom. Then, as is typical of his Friday afternoon reveries, he is rudely interrupted, this time by his computer terminal beeping to notify him of new e-mail in his inbox. He reads the subject line and fears the worst. As he pores over the attached files, he realizes that his most horrific fears have been realized. The firm's biggest client, "Guldmen Sechs", is asking him to document a transaction that makes the Travelers--Salomon Brothers acquisition/merger look like the purchase of a pack of chewing gum. As always, investment bankers' sense of time seems to be governed only by their "needs", without any reference to physical laws of nature. Apparently, the economy of our great nation, and indeed the world, will collapse if the documents are not on their desks when they come in Monday morning.This is a neat story, but do we have the "Right Stuff"? This may be a question not about our intelligence, but about the legal domain: Is it ripe for the advent of computer-based legal counselors? Well, when it comes to expert systems you have the right stuff when two general points ring true:A year ago, the attorney would have picked up the telephone and cancelled his weekend plans. The revised weekend "plan" would have been to live in the office with a team of two junior associates, two secretaries and one or two paralegals. But today, the attorney doesn't pick up the phone to cancel his weekend plans, nor to mobilize the troops. He looks at his clock again -- 4:20 -- rolls his sleeves back up and gets to work.
He begins an interactive conversation with his terminal, working with the artificially intelligent agent in much the same manner that he might instruct a junior associate. He tells the agent a little bit about the nature of the transaction and its high-level structure. The agent's embedded expertise begins to draft the requisite documentation even as it continues to interact with the attorney. As new information comes in, the documentation becomes increasingly refined and tightly tailored to the particularities of this transaction. Unlike the human junior associate, the artificial agent will not make careless mistakes in spelling or grammar, and will not forget a single item on the checklist of requisite legal provisions. Furthermore, the artificial agent will ensure that all of the documentation is legally consistent and complete. Finally, the agent produces a model of the documentation for you to review. Add serveral final details, have the agent draft a cover email for the atached document files and send it back over the ether to the investment banker. Time ... 5:00 ... shut the lights and out the door.
Two blocks away in his office high atop the World Trade Center, the investment banker put on his coat and hears his terminal beep. He innocently checks to see who has sent him e-mail. Bewildered at how the attorney was able to turn around the documents so quickly, he slumps down in his ergonomic chair and picks up the phone to cancel his weekend plans.
Of course, even when the domain seems appropriate, external considerations can discourage the use (and subsequent development of) such systems. For example, the medical domain seems to have the right stuff, but the public (patients) don't seem ready to accept the notion that a machine should assist a doctor not in the measurement of some vital statistic, but in the formulation of a medical opinion. Perhaps the gut reaction is that if consultation with a machine is required, the doctor must not know what he or she is doing (much the same way that a patient about to undergo surgery might feel somewhat uncomfortable if he saw his surgeon studying up on the procedure in a book minutes before - or during - the operation).
Our intuition is that people won't mind in the least if their lawyers are empowered by artificial colleagues--quite the contrary, if they are the beneficiaries of quicker turnaround time, lower legal fees and higher quality work product. If the case is won, if the deal is done, complaints will be few and far between. At least so it seems to us. Indeed, we believe that use of such systems will become de rigeur in law offices across the country, as ubiquitous as the telephone, often demanded by the clients. The case for Law as the perfect AI domain.
Allan Silver is president of Legal Knowledge Systems, an enterprise formed in 1995 to design and develop sophisticated software systems for the legal industry. Among a variety of software offerings, Legal Knowledge Systems produces a constraint-based, interactive configuration system which assists attorneys in the drafting of complex legal contracts. Prior to this venture, Mr. Silver was a practicing attorney at a large Wall Street law firm, specializing in international project finance. He has led teams of lawyers in the representation of commercial and investment banks, governmental lenders, private developers, and utilities in a variety of projects, including power generation, mining, telecommunications, papermill, and oil and gas projects. Before becoming involved in project finance, he worked in the legal department of a German commercial bank, and before beginning his legal career, he worked as an account executive in the commercial financing and factoring industry, supervising a multimillion dollar loan portfolio. Mr. Silver received a B.S. degree in Economics and an M.B.A. in Finance from New York University and his J.D. degree from Fordham University. He is co-author of an article entitled "Thais Welcome IPPs," (with G. Wigmore) which appeared in the September 3, 1993 issue of Project Finance International and an article entitled "In Support of Private Power," (with J. Green) which appeared in the Winter 1993 issue of Infrastructure Finance. Mr. Silver was a panelist speaking on state support at the September 1993 EXNET conference on Financing International Power Projects.
Selmer Bringsjord specializes in the logico-mathematical and philosophical foundations of artificial intelligence (AI), and, on the applied side, in the intersection of AI and creativity. He received his bachelor's degree from the University of Pennsylvania, and the PhD in philosophy and logic from Brown University in 1987. Since then he has been on faculty at Rensselaer, where in the Departments of Philosophy, Psychology and Cognitive Science, and Computer Science, he teaches AI, logic, and philosophy of mind. His pedagogy is in large part computation-based: All of his courses make intensive use of the Web, and of courseware of various types (e.g., Hyperproof). The materials thereon for his courses Introduction to Logic and Computability and Logic are in particular demand; they are used by publishers of logic courseware (e.g., Cambridge University Press and Stanford's CSLI). Bringsjord was on Rensselaer's team that won the prestigious Hesburgh Award (1995) for excellence in undergraduate education (for technology-based interactive learning). He was also a Lilly Fellow in 1989, during which time he designed and implemented an electronic textbook for introducing cognitive and computer science. He is co-director, with David Porush, of the Creative Agents Project, which has its roots in a project known as Autopoeisis, launched by a generous gift of $300,000 from the Luce Foundation and grants from Apple Computer. Bringsjord is author of the critically acclaimed What Robots Can \& Can't Be (1992, Kluwer; ISBN 0-7923-1662-2), which is concerned with the future of attempts to create robots that behave as humans. Two new technical books, Super-Minds, and Artificial Intelligence and Literary Creativity, are forthcoming this year (Kluwer Academic/Lawrence Erlbaum). The book Abortion: A Dialogue will also be published this Fall by Hackett. Dr. Bringsjord is also the author of a novel (Soft Wars}; Penguin, 1991), and papers ranging in approach from the mathematical to the informal, and covering such areas as AI, logic, natural theology, and ethics. He has lectured and interviewed in person and on radio and television across the United States, and in England, France, Ireland, Australia, Germany, Thailand, Japan and Canada.
Aspiring lawyers in the United States must take the LSAT (Law School Admissions Test) in order to enter Law School. We are building an artificial agent capable of taking (and scoring high on) the LSAT. (If it's true, as many hold, that lawyerly talent is required to excel on this test, then our agent will provide us with a foundation from which to build toward various technologies in the intersection of logic, AI, and Law.) The LSAT is divided up into different areas; for example, Logical Reasoning, Reading Comprehension, and Logic Games. We are tackling the Logic Games section first. (This section is sometimes called 'Analytical Reasoning.') Here are some sample problems from this section: