How machines reveal the gaps in evidence law

This short essay argues that the Federal Rules of Evidence contain critical gaps in terms of empowering litigants to meaningfully challenge the credibility of evidence. Specifically, the increasing use of machine-generated proof has made clear that evidence law does not offer sufficiently meaningful opportunities to scrutinize conveyances of information whose flaws cannot be exposed through cross-examination. These underscrutinized conveyances include machine-generated output, information conveyed by animals, and statements made by absent hearsay declarants. Even for some witnesses who can be cross-examined — such as eyewitnesses offering identifications and experts using a testable method — evidence law too often fails to subject their claims to meaningful scrutiny because of its overreliance on cross-examination.

As I explain below, these gaps have not always existed. That is, the rules of evidence have not always myopically focused on cross-examination as the primary means of testing human assertions, nor have they always excluded claims whose flaws cannot easily be tested by cross-examination from the scope of testimonial safeguards. Instead, this narrowing of evidence law appears to correspond with the rise of the “lawyerly art” of cross-examination in the mid-nineteenth century and a greater focus on the one testimonial infirmity of humans — insincerity — particularly suited to testing by cross-examination. Ironically, the legal case for a right to meaningfully scrutinize machine-generated proof might have been easier to make in 1823 than in 2023.

Of course, it is not too late to correct course. After explaining how evidence law has critical gaps because of an overemphasis on cross-examination and the danger of insincerity, this essay suggests changes to evidence rules and constitutional evidence doctrines that would better achieve accuracy — especially as proof becomes exceedingly technologically complex over the next fifty years.