Extraterritorial Jurisdiction: Can RICO Protect Human Rights? A Computer Analysis of a Semi-Determinate Legal Question
Engle, Eric Allen, The Journal of High Technology Law
I. USING COMPUTERS TO SIMULATE LEGAL DECISION MAKING
The ability of computers to perform complex tasks no longer remains a subject of serious debate. Thus, the question whether computers can model or simulate legal decision-making should be increasingly replaced by the questions: 1) how should computers assist in modelling legal decision-making and 2) what types of legal decisions are scientifically most interesting and useful to model. This article contends that semi-determinate "partially solved" legal problems represent the most interesting questions susceptible to computer analysis. A computer program accompanying this article demonstrates this proposition, by modeling an as yet undetermined question: whether the civil provisions of the Racketeering Influenced and Corrupt Organizations Act (RICO) (2) have extraterritorial affect.
As evidence of the general acceptance of computer decision-making ability, consider the development of tests for machine intelligence. This history demonstrates the capacity and limits of machine intelligence and which types of legal problems computer programs analyze most efficiently. Discussion of several of the more significant tests follows.
One of the first tests for artificial intelligence (A.I.), proposed by computer pioneer Alan Turing, (3) stated that machine intelligence becomes meaningful when a human no longer can distinguish the machine intelligence from human intelligence. (4) This test, known as the Turing Test, appears increasingly quaint. Although the Turing Test yields practical results, demonstrating that A.I. has evolved to a level of sophistication where humans cannot distinguish between responses generated by humans and those generated by artifical intelligence, it does not demonstrate that a computer can "think" in a cognitive fashion. (5) Critics of the Turing Test argue that it merely tests a computers ability to imitate human thought by applying a set of procedural rules. (6)
One of the most well known attempts at using A.I. to simulate human language in an electronic conversation was Joseph Weizenbaum's Eliza, a computer program designed to operate as a Rogerian therapist, which easily fooled many users into believing that they were talking online with a human psychotherapist. (7) Chat programs, however, have evolved further since Eliza and demonstrate increasing sophistication in their attempts to mimic human personality, including animation and speech synthesis. These programs still revolve around the concept of mirroring the human's input, though they also now employ algorithms to learn about the human, and sometimes even use distributed computing, (8) enabling them to grow beyond a sort of Lacanian "mirror stage" (9) and actually develop independent cognitive processes.
Once the simplicity of the Turing test became apparent, measures of A.I. shifted to more complex tasks and researchers developed better tests for A.I.. Another early test for A.I., which Turing worked on, asked whether the machine could play a competitive game of chess. (10) Rather than seeking to solve a general problem, namely creating the ability to mimic humans, researchers now directed computer intelligence towards solving precise, specific problems. This test proved a far more successful inquiry. (11) Contemporary chess programs can improve the humility of most of us by gently reminding us of our intellectual weaknesses. (12) As humans establish tests for machine intelligence, they program machines to meet those tests, pushing back the horizon of the question whether a machine can "think". (13) This cycle may even be inevitable: defining a problem is the first step to solving it so perhaps any definition of intelligence will (eventually) be programmable.
One test of A.I. not yet met, however, asks whether the computer possesses sentience, or self-awareness. To date, humans have not yet developed machines possessing self-awareness. But what is self-awareness and how can we recognize it? …