Eye on Eagle

Factual Background: A man was found alone and strangled in his apartment with multiple belongings missing. Multiple people claimed that Mr. Wakefield confessed to the murder and had been witnessed trying to sell the decedent’s belongings. Police arrested Mr. Wakefield and tested his DNA against multiple DNA mixtures present on the decedent and the cord with which he had been strangled. That data was sent to a company that used a continuous probabilistic genotyping software program called TrueAllele Casework System (“TrueAllele”). TrueAllele determined that it was millions, billions, and, in one case, quintillions times more probable that Mr. Wakefield, rather than an unrelated Black person, contributed to the DNA mixtures tested.

Mr. Wakefield moved for supplemental discovery disclosing the source code used by TrueAllele—a request rejected by the prosecution—and requested a Frye hearing. After the Frye hearing, the trial court held that the evidence put on by the prosecution showed that the TrueAllele system used methodology generally accepted within the relevant scientific community and ruled that there was no significant evidence to the contrary. While Mr. Wakefield objected that he was impaired at the Frye hearing because the TrueAllele source code was never provided to him, the trial court found that this claim was waived because Mr. Wakefield went forward with the Frye hearing without making that objection beforehand.

The Appellate Division, Third Department, affirmed. It found that the trial court was proper in finding that TrueAllele, which had been used and found admissible in various states, was scientifically supported. Further, it found that Mr. Wakefield’s right to confront witnesses was not violated by not having access to the TrueAllele source code, because source code itself does not qualify as an out-of-court declarant.

Issues Before the Court: 1) Whether the trial court abused its discretion by finding the TrueAllele DNA methodology to be “generally accepted” under the Frye standard. 2) Whether the non-disclosure of the TrueAllele source code violated Mr. Wakefield’s rights under the discovery law or under the confrontation clause.

Held: 1) The trial court properly found that the TrueAllele software used a system recognized by the relevant scientific community to yield reliable results and the source code itself was not needed to make this determination. 2) Although TrueAllele’s system involves artificial intelligence, failure to disclose the source code did not violate the confrontation clause because source code is not testimonial and, because it cannot be cross-examined, is not a declarant. Nor was it required under CPL § 240.40 because Mr. Wakefield failed to demonstrate a particularized need for the source code by motion to the court.

Judge Rivera, along with Judges Wilson and Troutman, concurred in the result, writing that the trial court erred by not ordering the disclosure of the TrueAllele source code so that the defense could provide third-party evaluation of it; however, the concurrence found this to be harmless error.

CAL Observes: In recent years, several new programs have been developed that conduct analysis that, a decade ago, would likely have been done by a human. While the reliability of these analyses may warrant Frye hearing, the majority opinion reinforces that growing prevalence and acceptability of such programs. That said, the 4-3 split regarding the importance and relevance of the source code of such programs may hint at further litigation exploring how to protect people’s right to cross-examination as the usage of artificial intelligence in prosecutions becomes more prevalent.

It also bears note that this decision leaves open the possibility—not preserved here—that the source code for a DNA program might need to be turned over when a defendant demonstrates a particularized need for the source code in order to exercise their right to cross-examine an analyst at a Frye hearing.