TL;DR:
- In May 2026 the Advisory Committee on Evidence Rules decided not to advance Proposed Federal Rule of Evidence 707, which would govern AI-generated evidence offered without an accompanying expert. The move delays any new nationwide gatekeeping standard for machine-generated conclusions until at least Fall 2026, pending further discussion and possible republishment.
- For trial teams, this means federal courts will continue to apply the existing Rule 702 and Daubert-style reliability standards to AI outputs, with no extra presidential push from a new AI-specific rule in the near term. Expect ongoing debates about how to handle AI-generated material in discovery, deposition, and trial, and plan for how to address AI outputs in evidentiary hearings using current frameworks.
- Practical readiness now centers on robust foundation, verification, and compelling objections when AI-generated material is offered without a live expert. Objection Academy’s objection drills and trial-readiness training remain a valuable way to shore up objections to AI-derived evidence, authenticate machine outputs, and prepare cross-examination for AI-driven conclusions.
What happened in May 2026
The federal government’s Advisory Committee on Evidence Rules held its May 2026 session amid intense public interest in whether AI-generated evidence should receive a dedicated gatekeeping rule. Public comment on Proposed Rule 707 had closed on February 16, 2026, after hearings and written submissions. By mid-May, the committee had not announced an adoption of a final version for the Standing Committee or Congress. A Bloomberg Law analysis published May 15, 2026 explained that the committee “decided not to advance to the next stage of the approval process a proposed rule governing the admissibility of AI-generated opinion-like evidence,” signaling a pause or potential overhaul rather than a straightforward adoption. This placed the rule on a longer timeline, potentially delaying any December 2027 effective date, assuming ultimate approval by the Supreme Court and Congress. (news.bloomberglaw.com)
The Committee’s May 2026 materials further reflect a nuanced debate about Rule 707’s scope and structure. The May 7 2026 agenda book and the May 17, 2026 Committee report show that the panel considered multiple versions of revised text, contemplated a notice framework, and discussed whether to republish a revised version for further public comment or to defer consideration to the Fall 2026 meeting. The discussions acknowledged mixed public comments and concerns from stakeholders, including calls to narrow the rule to machine-learning outputs and to address definitional boundaries such as “machine-generated evidence” versus routine or judicially noticeable technologies. (uscourts.gov)
For context, the agenda and report materials describe why Rule 707 had generated such debate. The committee noted that public feedback highlighted potential overbreadth and the risk of creating a de facto shortcut for admitting AI outputs without robust validation. At the same time, there was recognition that AI-generated material may already be appearing in federal cases and that some commentators favored moving forward with revisions that would provide clear reliability standards tied to Rule 702. The May 2026 materials therefore contemplated several versions and potential ways to proceed, including integrating AI-related reliability concerns with, or alongside, Rule 901 and 902 discussions or hosting a fall mini-conference on AI and deepfakes. (uscourts.gov)
Practical implications for trial teams
- Status quo remains: No new nationwide FRE 707 gatekeeping in force. Trial teams must continue to evaluate AI-generated material under the existing Rule 702/Daubert framework when offered without an expert witness. This means the proponent bears the burden to show that the AI output is based on sufficient facts, reliable methods, and a reliable application to the case, as under current 702(a)-(d). The absence of a final FRE 707 means fewer formal procedural hurdles dedicated to AI outputs, but more emphasis on traditional reliability testing and cross-examination for AI-derived conclusions. (news.bloomberglaw.com)
- Expect ongoing strategic debates: While a dedicated rule is paused, courts and practitioners continue to navigate questions about AI in discovery, privilege, and protective orders. Several firms have already noted AI-specific protective-order provisions in early court orders, signaling risk management as a live practice area even before a formal FRE 707 framework is approved. Litigators should monitor developments and consider how to tailor protective orders and disclosure protocols when AI tools are used in preparation. (kirkland.com)
- Preparation and objections remain essential: Without a finalized FRE 707, counsel should emphasize robust foundation, validation, and cross-examination practices for AI outputs under existing standards. This includes challenging the reliability of outputs that would be treated as expert conclusions if offered without a sponsoring expert, and preparing to resist AI-generated assertions that exceed permissible lay or non-expert testimony. Objection training that focuses on AI-driven evidence, authentication, and the reliability chain will remain highly practical for trial teams. Objection Academy’s drills can help teams rehearse these exact objections and refine courtroom performance when AI outputs surface in hearings or trials.
How to prepare now
- Maintain rigorous evidence foundations: Continue to treat AI-generated outputs as potential scientific or technical material that must meet the standards of Rule 702 if offered without a sponsor, or as admissible only with appropriate foundation under Rule 901/902 where applicable. Prepare to demonstrate data authenticity, source reliability, and method validity in each case.
- Build a practical AI readiness playbook: Develop a memo or checklist for cases likely to involve AI-derived materials, detailing how to verify outputs, how to lay foundation for the jury, and how to frame cross-examination around AI limitations.
- Train with focused practice: Use objection drills to rehearse objections to AI-driven conclusions, requests for admission of AI-generated materials, and cross-examination strategies for AI-produced evidence. Objection Academy’s tools can help trial teams simulate court-room objections and refine response timing and framing.
- Watch for Fall 2026 developments: The May 2026 materials indicate the possibility of republishing revised Rule 707 text for public comment or deferring to the Fall 2026 meeting. A fall mini-conference on AI rules is possible, which could lead to a revised path forward. Stay alert for formal updates from the Advisory Committee or the Courts press office. (uscourts.gov)
Next steps and timeline
Fall 2026 reconsideration: The May 2026 discussions left open the possibility of republishing a revised Rule 707 for another round of public comment or folding the issue into a broader AI rules discussion in Fall 2026. The agenda books and committee reports confirm ongoing consideration of AI-related gatekeeping and a potential mini-conference on the topic. Practitioners should track the Standing Committee’s agenda and any republication notices. (uscourts.gov)
Possible future alignment with other AI rules: As AI governance evolves, Rule 707 may converge with discussions around deepfakes, Rule 901/902 authentication, and Rule 702 reliability standards. While this may eventually yield a unified approach to AI evidence, the May 2026 pause underscores that any final framework will require careful drafting and broad consensus among practitioners, judges, and policymakers. (uscourts.gov)
Practical takeaway for trial teams: In the interim, prepare for both futures—continue to rely on robust traditional foundations for AI-generated evidence and use targeted objections where appropriate. Maintain discovery protocols that address machine-generated materials, and train litigation teams to respond quickly to AI-based assertions in motions, depositions, and trial.
Objection Academy remains a valuable ally for trial teams seeking to sharpen readiness for AI-involved proceedings. By focusing on reliable objections, authenticating machine outputs, and rehearsing courtroom confrontations, litigators can stay ahead of evolving evidentiary standards even as the federal rules of evidence await final, unified direction on AI.
Sources:
- Bloomberg Law, ANALYSIS: Proposed FRE 707 Stalls—But Is That Such a Bad Thing? May 15, 2026.
- Advisory Committee on Evidence Rules, May 2026 materials, including the May 17, 2026 committee report and the broader May 2026 agenda book.
- Advisory Committee on Evidence Rules, May 2026 Agenda Book (released May 7, 2026).
- United States Courts, Advisory Committee on Evidence Rules – May 2026 (Committee Reports).
- United States Courts, Evidence – May 2026 (Agenda Books).