Proposed New Federal Rule Regarding AI-Generated Evidence

Evidence generated by artificial intelligence (“AI”) is making its way into courtrooms. As a result, the U.S. Judicial Conference’s Advisory Committee has proposed a new rule “to regulate the admissibility of machine evidence that is introduced without the testimony of any expert.” The proposed rule marks an important effort to address concerns raised by the use of AI in courts of law and has generated extensive discussion and comments from judges, lawyers, and other interested parties across the country.

As proposed by the Advisory Committee, Federal Rule of Evidence 707 (“Proposed Rule 707”) would subject “machine-generated evidence” to the same admissibility standard as expert testimony. The proposed rule reads:

When machine-generated evidence is offered without an expert witness and would be subject to Rule 702 if testified to by a witness, the court may admit the evidence only if it satisfies the requirements of Rule 702 (a)-(d). This rule does not apply to the output of simple scientific instruments.

The Advisory Committee released Proposed Rule 707 for public comment in August 2025, and the public comment period closed this week on February 16, 2026.

Select issues raised by commenters in response to Proposed Rule 707 include:

• The phrase “machine-generated evidence” is undefined. While the Committee Notes illustrate that this rule is intended to address AI-generated evidence, the absence of a definition of those terms creates vagueness and ambiguity, and the scope of evidence targeted by the rule is unclear.
• The procedure for admitting machine-generated evidence using Rule 702 standards is uncertain because there might not be a qualified expert to examine regarding the proffered evidence. Thus, evidence offered under Proposed Rule 707 could spur more disputes and higher litigation costs than traditional expert testimony.
• Proposed Rule 707 may be premature at this time. AI technology and the public’s understanding of such technology continues to evolve rapidly, and it is unclear whether this rule creates the flexible framework needed to provide guardrails for such technology efficiently.

Notably, Proposed Rule 707 only attempts to address concerns with admissibility of evidence a party acknowledges to be AI-generated, and does not begin to address problems with unacknowledged AI-generated evidence, like deepfakes. Amendments and additions to Federal Rule of Evidence 901 have been suggested to deal with such authentication issues, but the scheme the Advisory Committee will choose to deal with the myriad problems generated by AI in the courtroom remains to be seen.

The Advisory Committee will review and process the public comments on Proposed Rule 707 and provide a final report on the matter in June 2026.