The Turing Test and the Legal Process
University of Central Florida; Pontifical Catholic University of Puerto Rico
Information & Communications Technology Law, Vol. 21, No. 2, June 2012, pp. 113-126
Michael Mauboussin: Here’s what active managers can do
The author proposes a novel thought-experiment, the “Turing litigation game.” Specifically, we propose replacing the existing arcane and archaic systems of civil and criminal procedure with a simple and probabilistic litigation game resembling the Turing Test from the world of computer science. The remainder of this paper is thus divided in four parts. Following this brief introduction, Part 1 provides some background by describing the original Turing Test and explaining how the Turing Test resembles the process of adjudication. Part 2 then describes our proposed Turing litigation game and identifies the conditions for implementing this alternative approach to litigation, while Part 3 introduces the possibility of probabilistic verdicts (as opposed to the traditional system of binary verdicts). Part 4 concludes.
The Turing Test and the Legal Process – Introduction
If we could design and develop a more rationale and less-error prone system of adjudication, what would it look like? Ideally, if we could neglect the constraints of custom, path dependence, and politics, we would “start from scratch” and propose an alternative and totally new system of civil and criminal adjudication. Specifically, we would propose a simplified and openly probabilistic procedural system resembling the “Turing Test” or “Imitation Game”, a well-known test of artificial intelligence proposed by the computer scientist Alan Turing in 1950. The Turing Test has generated extensive commentary in the fields of computer science, artificial intelligence, and the philosophy of mind,3 and as we explain in greater detail below, the logic of the Turing Test is also relevant to the legal process.
Although the Turing Test was designed to test for the intelligence of computer programs—that is, to answer the general question, “Can machines think?”-and was not designed to determine a particular defendant’s civil or criminal liability, we would extend the Turing Test to law and apply a Turing-like procedure to civil and criminal proceedings, but instead of testing for intelligence, our system would „test? for civil and criminal liability.
In his famous 1950 paper, the computer scientist Alan Turing described a simple question-and-answer game, what he calls the “imitation game”, involving three players: a man (player A), a woman (player B), and an interrogator (player C), who may be of either gender. In summary, the interrogator is allowed to put questions in writing to players A and B, and based on the responses provided by A and B, the interrogator must guess their true identities, or in Turing?s words: „the object of the game for the interrogator is to determine which of the other two is the man and which is the woman?, while „the object of the game for third player [i.e., player B, the woman] is to help the interrogator?.
Notice, then, that the object of the other player in Turing?s game—player A, the man—is to deceive or fool the interrogator about the truth of his gender (and about the truth of the other player?s gender as well). Next, Turing proposes that the role of player A be played by a digital computer: „What will happen when a machine takes the part of A in this game?? According to the “standard interpretation” of the Turing Test, the role of the interrogator is to determine which of the players is the person, and which is the computer.
Superficially, Turing?s imitation game appears to be totally different from the process of litigation in several ways. For instance, unlike most forms of civil and criminal litigation, in which all the main players (i.e., judge, jury, lawyers, and parties) are present in the same room and communicate orally, in Turing?s game the players must remain in two separate rooms and must communicate with each other in writing, either electronically or through an intermediary. Also, Turing’s imitation game resembles more an inquisitorial system of adjudication than an adversarial system, since it is the interrogator himself (or herself)—and not the other players or their representatives—who formulates the questions and evaluates the responses in the imitation game.
See full PDF here.