WLF Counsels Rejection of Proposed New Federal Evidentiary Rule for AI

“Just like some of the technology the proposal is aimed at, Rule 707 is not ready for primetime.”
—Zac Morgan, WLF Senior Litigation Counsel

Click here to read WLF’s comment.

(Washington, DC)—Washington Legal Foundation (WLF) today urged the Advisory Committee on Evidence Rules to reject Proposed Rule 707, a draft evidentiary rule targeting AI-generated evidence. While applauding Rule 707’s goals, WLF argued that its text would likely encourage, rather than discourage, the introduction of unreliable and insufficient evidence at trial.

The Federal Rules of Evidence govern the introduction and use of evidence, including witness testimony, in the federal courts. Proposed Rule 707, which targets “machine-generated evidence” while excluding “simple scientific instruments,” is a response to the rapid rollout of probabilistic, preventive, generative computer programs—often called artificial intelligence (AI).

WLF’s comments focused on three defects in the proposal. First, “machine-generated” is a capacious term that includes non-AI technologies. Second, the Proposed Rule could be read to allow courts to introduce unreliable AI outputs where it’s merely “more likely than not” that the program isn’t “hallucinating.” Third, the exclusion of “simple scientific instruments” is poorly defined and risks undermining other rules designed to keep junk science out of the courtroom.