Absolute truth or deus ex machina? The legal and philosophical ramifications of guilt-assessment technology.

• Author: Stronge, Aaron M.

Cite as 10 J. HIGH TECH. L. 113 (2009)

"[Y]ou must understand, sir, that a person is either with this court or he must be counted against it, there be no road between. This is a sharp time, now, a precise time-we live no longer in the dusky afternoon when evil mixed itself with good and befuddled the world. Now, by God's grace, the shining sun is up, and them that fear not light will surely praise it."

- Arthur Miller, The Crucible, 76 (Heinemann 1992) (1953).

1. Introduction

   On June 12, 2008, a court in India convicted Aditi Sharma of murdering her fiance and sentenced her to life in prison. (1) As its primary evidence against her, the court relied on a brain scan test that purported to show that Sharma's brain held knowledge of the crime that only the killer's brain could contain. (2) Researchers in the United States have expressed strong opposition to this verdict, noting that the technology employed is far from perfect and should not have been relied upon in this manner. (3)

   Since the first vestiges of societal establishment, we have struggled with not only how to punish those who do wrong, but how to determine the identity of the wrongdoer. (4) In the earliest Anglo-American trials, the relied upon means of assessing the guilt of the accused was not a trial by his peers, but one under God. (5) Under judicium Dei, the accused was either subjected to some form of torture, as in trial by ordeal, or to trial by combat. (6) These two forms of "trial" shared the same inspiration: God would not let an innocent suffer, and if the accused were truly not guilty, He would intervene on his behalf. (7)

   The commonality that judicium Dei shares with modern attempts at guilt assessment by technological means is that the responsibility for deciding the accused's fate is placed not upon society, but rather on some third party--in the latter case, a machine. Ben Clark, of the Notre Dame University of Australia School of Law, wonders whether in the creation of lie detection technology we have not merely "invented a modern form of witch dunking." (8)

   However, although in its present state such technology is not foolproof--claims of accuracy at detecting deception hover around ninety percent (9)--the benefits cannot be ignored. Laboratory tests have shown people to be very poor lie detectors, with success rates averaging that of pure chance. (10) Thus on its face, with the primary purpose of our legal system to separate the guilty from the innocent, any device that allows us to do so with better accuracy should be embraced. In actuality, the issue is far more complicated, and many fear that the general acceptance of lie detection evidence-or that of any other "mind reading" technology-will unleash an Orwellian nightmare. Is it possible to utilize this technology to achieve the most accurate fact-finding possible without completely usurping the role of the jury and the ideals for which the jury system stands?

   In Part II of this article, I will explore the history of scientific evidence generally, as well as how the admission standards of the judicial system have adapted with the introduction of new technologies. I will then look more specifically at the history of lie detection, including emerging neuroscience technologies and how they have been both employed and dismissed by the courts. In Part III, I will broadly articulate the areas in which the use of neuroscience technologies conflicts or raises issues with the judicial process, particularly with respect to the role of the jury. Finally, in Part IV, I will examine the pros and cons of judicial utilization of these technologies, as well as the broader philosophical implications of doing so.

2. History

   1. Admission of scientific evidence generally

      Scientific evidence can provide a solid foundation for proving a given point in a case, thereby reducing the variables that must be considered by the jury and lessening the possibility of error. For the majority of the 20th century, the admission of scientific evidence was governed by the rule established in Frye v. United States. (11) Frye, decided in 1923, put forth the standard that in order for scientific evidence to be admissible in a court of law, the method by which that evidence was procured must be "sufficiently established to have gained general acceptance in the particular field in which it belongs." (12)

      In 1976, the Federal Rules of Evidence (FRE) (13) were adopted, taking a far more liberal stance on what scientific evidence could be admitted. (14) Specifically, Rules 702 and 703 consolidated admission requirements for scientific evidence and expert testimony by asking: (1) whether or not the evidence would aid in settling a certain question of fact; (2) whether the witness had sufficient expertise in his area; and (3) if the data that the witness were relying upon would not be admissible alone, if that data could be reasonably relied upon by experts in the field. (15)

      In 1993, in its ruling of Daubert v. Merrell Dow Pharmaceuticals, Inc., (16) the Supreme Court acknowledged and sought to reconcile the polar admissibility standards of Frye and the FRE. (17) To that end, the Court established a two-pronged analysis for judges to employ in determining whether to admit scientific evidence or expert testimony. (18) First, the judge must decide whether the expert's testimony is in fact to be considered "scientific knowledge," whether the findings themselves are "derived by the scientific method," and whether the work can be considered "good science." (19) The second prong, which the Court referred to as the "fit" requirement, obliges proffered testimony to be relevant to a material portion of the advancing party's case. (20)

      In addition to the two-pronged test, the Court also listed several factors to provide further guidelines for whether or not a judge should admit expert scientific testimony under Rule 702. (21) These factors included: whether or not the methodology employed by the expert is generally accepted by the scientific community, whether that methodology has been subjected to peer review and publication, the degree to which it has been tested, and whether the associated rate of error is acceptable. (22)

      The task before judges in their post-Daubert "gatekeeping role" (23) is a difficult one. Not only must they must decide whether or not scientific evidence should be admitted, but also who has the necessary expertise to testify about that evidence and what that testimony may entail. (24) Further, and perhaps most importantly, the judge must also weigh the prejudicial effect the evidence could have upon the jury against the probative value of that evidence. (25) The complexities involved in deciding whether scientific evidence meets admissibility standards may be evinced in the courts' ongoing and tumultuous relationship with lie detection and brain scan evidence.

   2. Admission of Lie Detection Evidence

      Historically, brain scan evidence has proven very persuasive to juries in establishing the degree of responsibility of criminals. (26) At John Hinckley's trial for his assassination attempt on President Reagan in 1981, the court allowed the presentation to the jury of computerized axial tomography (CT) scans of Hinckley's brain, as well as the testimony of a psychiatrist who noted the presence of cerebral atrophy. (27) Despite the lack of scientific evidence linking the atrophy to Hinckley's actions, as well as testimony by a radiologist that the abnormality present in the scans could not reasonably be relied upon as the root cause of those actions, the jury still found Hinckley not guilty by reason of insanity. (28)

      Beyond evidence of brain pathology that could absolve a defendant of responsibility for his crimes, there exists the possibility of using more advanced brain scanning technology to determine whether or not that defendant actually committed the crime in question. (29) However, the utilization of neuroimaging techniques to enable lie detection presents certain problems, not the least of which being the dubious nature of the history of polygraph evidence in the United States. (30)

      1. Lie Detection by Polygraph

         The judicial system's turbulent relationship with lie detection devices began with the Frye court. (31) In addition to setting a standard for the admission of scientific evidence that remained predominant for seventy years, Frye's assessment of the first systolic blood pressure deception test also expressed a suspicion of lie detection devices that remains largely untouched today, (32) despite vast improvements in the technology. (33)

         In 1996, in United States v. Scheffer, (34) the United States Court of Appeals for the Armed Services reversed the conviction by general court-martial of airman Edward Scheffer on the grounds that the exclusion of polygraph evidence by the military judge was a violation of his Sixth Amendment right to present a defense. (35) Subsequently, however, the Supreme Court reversed the decision, holding that a per se exclusion of polygraph evidence did not violate the Sixth Amendment because such evidence is not in itself factual:

         The raw results of a polygraph exam--the subject's pulse, respiration, and perspiration rates--may be factual data, but these are not introduced at trial, and even if they were, they would not be "facts" about the alleged crime at hand. Rather, the evidence introduced is the expert opinion testimony of the polygrapher about whether the subject was truthful or deceptive in answering questions about the alleged crime. A per se rule excluding polygraph results therefore does not prevent an accused--just as it did not prevent [Scheffer] here--from introducing factual evidence or testimony about the crime itself .... (36) Following Scheffer, the issue of polygraph evidence admission was further addressed in United States v. Waters, (37) which involved a prosecution for aggravated sexual abuse of a child. (38) At trial, the defendant sought to...