|
Evidence - Artificial Intelligence. R. v. Prasad [necessity]
In R. v. Prasad (Ont CA, 2024) the Ontario Court of Appeal considers the 'necessity' element of the first step in the White Burgess expert evidence test, here in assessing whether 'expert' psychological evidence (about 'priming') is useful to aid the fact-finder in assessing their trust in "digitally enhanced audio" ('artificial intelligence'?) recordings:Necessity and the role of the expert in the trial process
[35] The more difficult question is whether the expert evidence is necessary to enable the trier of fact to draw accurate inferences from the tendered audio evidence. The cases have linked the criterion of necessity with concerns about the trial process and the danger that an expert might usurp the role of the trier of fact. In my view the expert evidence does not meet the test of necessity, as I will explain.
[36] The trial process, with its allocation of responsibilities between the judge and the trier of fact, is primordial: the trial process must be respected, particularly the role of the trier of fact, which must never be usurped by another. Consequently, to be admissible and to guard against such usurpation, opinion evidence must meet more exacting criteria.
[37] Justice Cromwell expressed the longstanding policy of the law, quoting Thayer, who said in 1898 that it is "for the jury to form opinions, and draw inferences and conclusions, and not for the witness": White Burgess, at para. 14. The point, added Cromwell J., “is to preserve trial by judge and jury, not devolve to trial by expert”: White Burgess, at para. 18. See also R. v. D.D., per Major J., at para. 49. While Cromwell J.’s quoted comments might appear related to the second White Burgess step of gatekeeping, I introduce them here because he noted that relevance, necessity, and reliability are “part of a sliding scale” that plays a role “in weighing the overall competing considerations in admitting the evidence”: White Burgess, at para. 25
[38] Justice Cromwell expressly approved, at para. 24 of White Burgess, Doherty J.A.’s statement in Abbey that the “trial judge must decide whether expert evidence that meets the preconditions to admissibility is sufficiently beneficial to the trial process to warrant its admission despite the potential harm to the trial process that may flow from the admission of the expert evidence” (emphasis added): Abbey, at para. 76.
[39] Justice Doherty cautioned: “the most important risk is the danger that a jury will be unable to make an effective and critical assessment of the [expert opinion] evidence” because of complexity and “impenetrable jargon”: Abbey, at para. 90. He warned that, in addition “to the risk that the jury will yield its fact-finding function, expert opinion evidence can also compromise the trial process by unduly protracting and complicating proceedings”: at para. 91.
[40] The necessity inquiry does not examine the expert evidence in isolation. It asks whether the expert evidence is needed to enable the trier of fact to perform its fact-finding function properly and accurately.
Questions or criteria to guide thinking about whether expert evidence is necessary
[41] The cases have generated considerations that can be framed in terms of questions or criteria for assessing whether expert evidence is necessary. I consider several.
[42] First, is the subject matter of the expert evidence within or beyond the common experience of lay people? Major J. adopted the words of Professor Paciocco that expert evidence should be admitted: “Only when lay persons are apt to come to a wrong conclusion without expert assistance, or where access to important information will be lost unless we borrow from the learning of experts”: R. v D.D.,[7] at para. 57.
[43] A useful distinction can be made between hard and soft sciences. Professor David Paciocco (as he then was) stated: “While distorting opinions can occur within the hard sciences, the risk that they will exist is more prevalent with behavioural science.” He explained: “Reliability problems, while not peculiar to behavioural science, tend to be more intense for the behaviourist than for the ‘hard scientist’”[8].
[44] Indeed, judges and juries might be tempted to defer to experts in the field of the behavioural sciences. Justice McLachlin (as she then was) observed in Marquard, at para. 51, that “expert evidence on human conduct and the psychological and physical factors which may lead to certain behaviour relevant to credibility, is admissible, provided the testimony goes beyond the ordinary experience of the trier of fact” (emphasis added). In R. v. McIntosh,[9] Finlayson J.A. cautioned, at para. 14: “We are too quick to say that a particular witness possesses special knowledge and experience going beyond that of the trier of fact without engaging in an analysis of the subject matter of that expertise.”
[45] The notional line between what is within the normal experience of a trier of fact might change over time and might vary between localities, as O’Connor J.A. pointed out in R. v. D.S.F.[10] This line must be continually reassessed.
[46] Second, and relatedly, how might the admission of the actual subject matter of the expert evidence affect the trial process? There are many cases in which expert psychological evidence – the product of a soft science – has been rejected on the basis that the testimony does not go beyond the ordinary experience of the trier of fact. In McIntosh, Finlayson J.A. set out the area on which the expert evidence was proposed to testify, at para. 11:Dr. Yarmey commented on the factors present at the time of the robbery that would impair the witnesses’ ability to make an accurate identification, the problem of cross-racial identification, the quality of memory recall for perceived events over different time spans, the influence of “post event information” on memory, the validity of the photographic lineup, the misconceptions of jurors with respect to photographic lineups, the difficulties with “in dock” identifications, and police procedures relating to the identification of the two accused persons. [47] Justice Finlayson found that the trial judge was right to refuse to admit the evidence because the expert “is not testifying to matters that are outside the normal experience of the trier of fact: he is reminding the jury of the normal experience” (emphasis added): at para. 20. See also R. v. Frimpong (refusing expert evidence on the frailty of eye-witness testimony),[11] and R. v. Pearce (refusing psychological evidence on false confessions).[12]
[48] There is plainly no rule banning evidence as to the psychology of a witness. Much depends on the context. In R. v. R.D., the Supreme Court held that the evidence of a psychologist tendered by the defence on the reliability of the memories of children was admissible, while in R. v. D.D., decided four years later, the Supreme Court held that similar evidence tendered by the Crown was not admissible.
[49] I would draw an additional distinction between evidence about how psychological influences might affect a witness, and evidence about how psychological influences might affect a judge or jury. In this case, the proposed evidence touches on both. The police officer giving evidence as to drug culture lingo and the police transcriptionist would both be affected by priming, as would the judge as the trier of fact, on the theory advanced by the experts.
[50] It seems to me that a fundamental challenge to the routine, longstanding, and time-tested operation of the system of justice, such as how psychological influences might affect a judge or jury, should be treated with great caution, if not serious skepticism. We upend the common law at our peril.
[51] The third question is this: does the expert evidence express an opinion on the very question that the trier of fact must answer? There is judicial reluctance to freely admit such opinion evidence because doing so might usurp the task of the trier of fact. I pick out here, once again, the words of McLachlin J. in Marquard: “there is a growing consensus that [...] expert evidence on the ultimate credibility of a witness is not admissible”: at para. 51. In Mohan, Sopinka J. called for stricter scrutiny: “The closer the evidence approaches an opinion on an ultimate issue, the stricter the application of this principle”: Mohan, at para. 32. See also R. v. J.(J.‑L.).[13]
[52] This problem that undue weight might be given to the expert opinion on the “very question” that the trier of fact must answer can be avoided by more careful instructing questions put to an expert by counsel who is alert to this problem. Sometimes it can be overcome by the trial judge ordering partial redactions of the text of expert reports. Ultimate reliability, along with credibility, is the province of the trier of fact. Of course, trial judges too must not permit themselves to be unduly influenced by the opinions of experts on ultimate questions.
[53] Fourth, could the caution implicit or explicit in the expert opinion be adequately communicated by a jury instruction or by a judge’s implicit or explicit self-instruction? For example, in R. v. D.D., as noted, the majority of the Supreme Court found that the psychological evidence should not have been admitted. Major J. said, at para. 58: “In my view, the content of the expert evidence admitted in this case was not unique or scientifically puzzling but was rather the proper subject for a simple jury instruction.” See also McIntosh, at paras. 22 and 26. Justice Doherty concluded in Abbey that, in “addressing the extent to which the opinion evidence is necessary, the trial judge will have regard to other facets of the trial process – such as the jury instruction – that may provide the jury with the tools necessary to adjudicate properly on the fact in issue without the assistance of expert evidence”: at para. 95. At paras 54-83 the court considers the psychological theory of 'priming' as it relates to 'filling in the blanks' in poor quality audio recording.
|
