*Dallon Danforth
“[T]he Confrontation Clause guarantees more than the right to ask questions of a live witness…” – William J. Brennan Jr., Associate Justice of the Supreme Court of the United States of America.[1]
I. Introduction
One day, for seemingly no reason at all, you are arrested and charged with the murder of a victim that you claim to have never met. As discovery commences, you learn that law enforcement, in the course of their investigation against you, performed DNA analysis on various samples collected from the scene of the crime. A laboratory analyst produced a report concluding that certain forensic samples are more than hundreds of billions of times more likely to contain your DNA than that of an unrelated party. Knowing that you face life without the possibility of parole, how might you challenge the analyst’s conclusions in court? Does your decision change if the analyst is not a human being, but rather a complex, probabilistic genotyping algorithm developed by a private software company?
As improbable as this scenario appears, it was this exact series of events that ultimately lead to the conviction of New York resident, John Wakefield.[2]
II. Case Study
On May 27, 2015, John Wakefield was found guilty of murder in the first degree. During Mr. Wakefield’s trial, the state introduced evidence of DNA analysis performed by TrueAllele, an artificial intelligence algorithm developed by Cybergenetics.[3] Mr. Wakefield’s defense was severely prejudiced by the results of the DNA analysis which estimated that certain DNA samples were up to 170 quintillion times more likely to belong to Mr. Wakefield than an unrelated person of identical racial background.[4]
Prior to his trial, Mr. Wakefield moved for receipt of TrueAllele’s “source code,” citing his right to confront the witnesses brought against him under the Sixth Amendment.[5] Mr. Wakefield argued that TrueAllele represented “the functional equivalent of a laboratory analyst,” and because the report generated by that laboratory analyst was inherently testimonial, Mr. Wakefield had a Sixth Amendment right to confront TrueAllele as an expert witness against him.[6] The trial court denied the motion, finding that TrueAllele’s report was not testimonial in nature, and that any concerns regarding TrueAllele’s reliability were curable by allowing Mr. Wakefield’s counsel to cross-examine TrueAllele’s Chief Scientist and Executive Officer, Dr. Mark Perlin.[7]
After such cross-examination, Mr. Wakefield argued that Dr. Perlin’s “surrogate testimony” failed to satisfy his Sixth Amendment right, once again asserting that TrueAllele, rather than Dr. Perlin, was the actual expert witness brought against him.[8] Without TrueAllele’s source code, Mr. Wakefield argued, defense counsel could not accurately determine the reliability of TrueAllele’s methodology and, by extension, the validity of the results contained within the highly prejudicial report.[9] Following his trial, Mr. Wakefield was sentenced to life imprisonment without the possibility of parole.[10]
Mr. Wakefield appealed his conviction, arguing that “his right to confront witnesses was violated by not having access to TrueAllele’s source code.”[11] The appellate court affirmed the conviction, noting that the “TrueAllele report was testimonial in nature since it was generated to assist the police and prosecutors,” but that TrueAllele itself was not a declarant such that the Sixth Amendment would be applicable to its algorithm.[12] Mr. Wakefield appealed once more to New York’s highest court, which affirmed both the ruling and the reasoning of the intermediate court.[13] In a concurring opinion, Judge Rivera differed from the majority by finding that Mr. Wakefield had a Sixth Amendment right to review TrueAllele’s source code, while also criticizing the trial court’s shielding of the source code as a trade secret based upon an earlier argument made by Dr. Perlin.[14] Nevertheless, Judge Rivera concluded that the error was harmless in light of the considerable evidence that otherwise implicated Mr. Wakefield’s guilt.[15]
Following his defeat in New York’s highest court, Mr. Wakefield filed for a writ of certiorari.[16] The Supreme Court of the United States denied his petition on November 21, 2022.[17]
III. Predictive Analysis
With the “current unprecedented advancements in artificial intelligence,” such as those contained in TrueAllele’s algorithm, Mr. Wakefield’s asserted right to review advanced algorithmic source codes will almost certainly be re-litigated by future defendants.[18] Afterall, the increased use of powerful algorithmic software continues to foster novel legal issues.[19] For example, Georgia resident Randal Reid was charged with felony theft after security cameras equipped with powerful facial recognition software falsely identified Mr. Reid as a participant of the criminal event.[20] Mr. Reid, who had never set foot in the charging jurisdiction, was jailed for nearly a week before law enforcement rescinded his arrest warrant after acknowledging the error in the algorithm’s facial identification process.[21] Had Mr. Reid’s case proceeded to trial, Mr. Wakefield’s failed appeals suggest that Mr. Reid would be denied access to the source codes utilized by the facial recognition technology used to implicate his guilt.[22]
By outlining the competing interests that arise when courts attempt to apply the Sixth Amendment to an artificial witness, Judge Rivera’s concurrence reveals a powerful dichotomy. On one hand, “[t]rade secrets can be used to protect almost everything related to software,” and algorithms like TrueAllele therefore enjoy considerable protections under the law.[23] However, shielding the source code of testimonial algorithms by appointing surrogate witnesses with a proprietary interest in that same source code does not produce the “objective, unbiased review” necessary for the proper administration of justice.[24] Because the right to confront and cross-examine witnesses is “an essential and fundamental requirement for the kind of fair trial which is this country’s constitutional goal,” the confrontation of advanced algorithms via surrogate testimony places at odds the unstoppable force of the Sixth Amendment with the immovable object of trade secret protections.[25] Of course, both interests, compelling as they are, cannot indefinitely prevail against the other, and courts will soon be tasked with appointing a victor. As New York’s intermediate court previously noted, “[g]iven the exponential growth of technologies such as artificial intelligence, to embrace the future we must assess, and perhaps reassess, the constitutional requirements . . . that arise where law and modern science collide.”[26]
IV. Conclusion
Given the rapid rate at which advanced software algorithms continue to permeate everyday life, such technology will soon strain the Sixth Amendment in an unprecedented manner. As future defendants come to relitigate Mr. Wakefield’s novel argument, presiding courts will further weigh trade secret protections against the authority of the Sixth Amendment’s Confrontation Clause.
*Dallon Danforth is a second-year student at the University of Baltimore School of Law. He also serves as the President of the Caroll Inn Chapter of the Phi Delta Phi International Legal Honor Society and is a teaching assistant for Professor Matthew Lindsay’s Constitutional Law course. He would like to thank the Law Review staff for the opportunity to publish this piece and his friends for their never-ending support.
[1] United States v. Owens, 484 U.S. 554, 572 (1988) (Brennan, J., dissenting).
[2] People v. Wakefield, 107 N.Y.S.3d 487, 491 (N.Y. App. Div. 2019) (“A jury trial was held, after which defendant was convicted of murder in the first degree and robbery in the first degree.”).
[3] Id. “[TrueAllele] . . . eliminat[es] all but the most rudimentary of human participation.” People v. Wakefield, 38 N.Y.3d 367, 387 (2022).
[4] “Specifically, TrueAllele concluded that it was 5.88 billion times more probable that defendant was a contributor to the mixture on the amplifier cord . . . that it was 170 quintillion times more probable that defendant was a contributor to the mixture on the outside rear shirt collar . . . that it was 303 billion times more probable that defendant was a contributor to the mixture on the outside front shirt collar . . . and that it was 56.1 million times more probable that defendant was a contributor to the mixture on the victim’s dorsal . . . . ” Wakefield, 38 N.Y.3d at 373.
[5] Id. at 378; see also Rod Dixon, When Efforts to Conceal May Actually Reveal: Whether First Amendment Protection of Encryption Source Code and the Open Source Movement Support Re-Drawing the Constitutional Line Between the First Amendment and Copyright, 1 Colum. Sci. & Tech. L. Rev. 3, 55 (2000) (“[S]ource code is the text of a program written in a ‘high-level’ programming language, such as ‘PASCAL’ or ‘C.’”).
[6] Wakefield, 38 N.Y.3d at 378.
[7] Id. at 374, 378.
[8] Id. at 378.
[9] See id.
[10] Id. at 394 (Rivera, J., concurring).
[11] People v. Wakefield, 175 A.D.3d 158, 165 (N.Y. App. Div. 2019) (acknowledging that “[t]his argument raises legitimate and substantial questions concerning due process as impacted by cutting-edge science.”).
[12] Id. at 168–69.
[13] Wakefield, 38 N.Y.3d at 386.
[14] Id. at 387.
[15] Id. at 394.
[16] Docket Search, Sup. Ct. of the U.S., https://www.supremecourt.gov/search.aspx?filename=/docket/docketfiles/html/public/22-5588.html (last visited Sep. 3, 2023).
[17] Id.
[18] “[T]he current unprecedented advancements in artificial intelligence devices constitute a sort of ‘Fourth Industrial Revolution’ . . . . ” Thomas Belcastro, Getting on Board with Roberts: How the Business Judgment Rule Should Apply to Artificial Intelligence Devices Serving as Members of a Corporate Board, 4 Geo. L. Tech. Rev. 263, 270 (2019).
[19] See, e.g., John Simerman, JPSO Used Facial Recognition Technology to Arrest a Man. The Tech was Wrong., nola (Jan. 2, 2023), https://www.nola.com/news/crime_police/jpso-used-facial-recognition-to-arrest-a-man-it-was-wrong/article_0818361a-8886-11ed-8119-93b98ecccc8d.html; see also Arjun Sha, 18 Examples of AI You’re Using in Daily Life in 2023, Beebom (Apr. 29, 2023, 11:51 AM), https://beebom.com/examples-of-artificial-intelligence/.
[20] See generally Simerman, supra note 19 (“The case highlights the pitfalls of a technology that more law enforcement agencies are adopting across the country, even as critics point to research showing bad matches at higher rates for some populations, including Black people and women.”).
[21] Id.
[22] Id. (the law enforcement agency responsible for Mr. Reid’s arrest utilizes facial recognition algorithms developed by Clearview AI and Morphotrak).
[23] “Trade secrets can be used to protect almost everything related to software.” E. Robert Yoches, Protection of Computer Software by Patents, Trade Secrets, and Trademarks, 22 Tort & Ins. L. J. 354, 358 (1987).
[24] People v. Wakefield, 38 N.Y.3d 367, 395 (2022) (Rivera, J., concurring).
[25] Pointer v. Texas, 380 U.S. 400, 405 (1965); see also Rebecca Wexler, Life, Liberty, and Trade Secrets: Intellectual Property in the Criminal Justice System, 70 Stan. L. Rev. 1343, 1429 (2018) (“A criminal trade secret privilege would almost certainly lead to overclaiming, abuse, and the exclusion of highly probative evidence; it would also project a message that the government values intellectual property holders more than those whose life or liberty is at stake.”).
[26] People v. Wakefield, 175 A.D.3d 158, 165 (N.Y. App. Div. 2019).
University of Baltimore Law Review 