//Two rules, one dated status
Rule 707 and Rule 901(c): AI Evidence Status
As of September 19, 2026, proposed Federal Rule of Evidence 707 is not in effect and draft Rule 901(c) sits in the Advisory Committee's agenda materials as a draft; both go before the Advisory Committee's mini conference on October 15, 2026, and Matt Aubin, CDFE, FBCI, keeps this page current.
By Matt Aubin, Founder, Southern Recon Agency and E3 Legacy Intel. CDFE, FBCI. Investigating since 2009. Published . Updated .
The December 1, 2027 effective date circulating online is outdated. Rule 707 was approved for publication for public comment, which is not adoption, and it was then deferred.
What the record says: the Advisory Committee on Evidence Rules revised the rule and deferred it for further study on May 7, 2026, and the Standing Committee reported in June 2026 that it does not recommend action at this time. There is no adopted rule and no effective date.
Status as of September 2026
Proposed Rule 707: not law, no effective date. Public comment closed February 16, 2026; the Advisory Committee on Evidence Rules revised and deferred it on May 7, 2026; the Standing Committee reported in June 2026 that it does not recommend action at this time.
Draft Rule 901(c): a draft in the committee's agenda materials, never published for comment; the committee decided in May 2026 that an amendment was not needed yet and kept the text ready.
Next milestone: both rules go before the Advisory Committee's mini conference with technical experts on October 15, 2026. Sources: the rules committee agenda books and the August 2025 preliminary draft on uscourts.gov.
The rules are moving targets. The reason to prepare now is not a deadline. It is that the standard they point to, Rule 702, is already the law.
Is Rule 707 in effect?
No. As of September 2026, proposed Federal Rule of Evidence 707 is a draft, not law, and it has no effective date. It was published for public comment from August 15, 2025 to February 16, 2026, drew more than seventy written comments and a public hearing in January 2026, and was then revised rather than advanced. At its May 7, 2026 meeting the Advisory Committee on Evidence Rules deferred it for further study, and the Standing Committee reported in June 2026 that it does not recommend action on the proposal at this time. The December 1, 2027 effective date still repeated across the web is outdated. This is general information, not legal advice.
What is machine-generated evidence?
Machine-generated evidence is output produced by a process or system rather than by a person: a risk score, a classification, an enhanced or reconstructed image, a predictive coding result, a forensic analytic, an AI generated summary of records. Proposed Rule 707 is aimed at that output when it is offered without a human expert to sponsor it and it would be treated as expert opinion if a person had said it. In that case the output would have to meet the Rule 702 reliability test. Simple instruments such as thermometers and scales would be carved out. Everything with judgment built into it would be in scope.
The proposed text is in the committee's preliminary draft of proposed amendments, August 2025 (PDF), and the test it borrows is Rule 702.
What is draft Rule 901(c) and why does it matter for deepfakes?
Draft Rule 901(c) would set out how a court handles evidence that a party claims was fabricated or altered by artificial intelligence, which is the deepfake question. As drafted it shifts the burden in two steps: the opponent must first offer enough for a reasonable person to infer that the evidence was fabricated, because a bare claim of deepfake is not enough, and only then must the proponent prove authenticity to a higher standard than Rule 901 normally requires. It matters because it is the rule that would govern a challenged recording, image, or message in court. It is at an earlier stage than Rule 707: it has never been published for public comment. The Advisory Committee studied it across six meetings, drafted text, and decided in May 2026 that an amendment was not needed yet, keeping the draft ready in case deepfake cases increase.
Today's authentication rule is Rule 901. The draft text and the committee's discussion of it are in the evidence rules agenda books on uscourts.gov.
Send the evidence and the deadline.
AI touched an exhibit and you need to know what it is worth before you build on it. Send it with the date you are up against.
What did the May 2026 and June 2026 committee decisions actually say?
On May 7, 2026 the Advisory Committee on Evidence Rules had both drafts in front of it and advanced neither. It revised the text of proposed Rule 707 and deferred it for further study together with the deepfake problem, and it kept draft Rule 901(c) as a study item rather than publishing it. In June 2026 the Committee on Rules of Practice and Procedure, the Standing Committee, reported that it does not recommend action on proposed Rule 707 at this time, that the proposal has been revised, and that study will continue. The next step for both is a mini conference with technical experts and practitioners at the Advisory Committee's fall meeting on October 15, 2026. Nothing is in effect and no effective date exists for either rule.
The primary record is the committee's own agenda books and reports on uscourts.gov, and the current Federal Rules of Evidence. This page is reviewed against them and the date at the top changes when it is.
What should counsel do now with AI assisted evidence?
Build the reliability record before anyone asks for it, on principles that hold no matter what the final rules say: sufficient and lawfully sourced inputs, a documented and repeatable method, the specific system and version recorded, independent human verification of every output, a chain of custody, and disclosure of how AI was used. That is the Evidence-Grade AI Standard, and it tracks Rule 702, which is already the law. If the other side's evidence is AI assisted or AI generated, demand the same record in discovery; the party that cannot produce it has an assertion, not a finding. A consulting expert behind counsel can build or test that record and write the analysis without being disclosed.
The framework is the Evidence-Grade AI Standard. The consulting seat is described on AI evidence consulting expert for attorneys, and a firm's whole workflow is validated by the AI Evidence Audit.
Does Rule 707 cover deepfakes?
Not directly. Rule 707 as drafted addresses evidence that is offered as machine generated. A disputed deepfake, where the fight is whether the media is real at all, is an authentication question, which is what draft Rule 901(c) would govern. Since May 2026 the committee has treated the two problems as one workstream, and both go to the October 15, 2026 mini conference. Either way, a challenged recording still needs a qualified examiner and, behind counsel, someone who can turn the examination into a discovery plan and a written analysis. That is the consulting expert role Matt Aubin fills.
How that works for the attorney: AI evidence consulting expert. Deepfake examinations themselves run through E3 Legacy Intel at e3intel.io/deepfake-verification.
//The record
Six things that carry an AI assisted finding through a reliability challenge
They meet what Rule 702 asks today, and they are built to survive what 707 and 901(c) are likely to require. They map directly to the Evidence-Grade AI Standard.
Sourced, lawful inputs
Every input the finding depends on is identified, preserved, and lawfully obtained under the rules that govern the work without AI.
A documented, repeatable method
The method is written down so another examiner could follow it, with the specific system and version recorded.
Independent human verification
A qualified human verifies every output against source evidence before it becomes a finding. Unverified AI output is a lead, not a conclusion.
Chain of custody and provenance
A record of what data, what system and version, what steps, who verified, and when, preserved to survive challenge.
Disclosure, calibrated
How AI was used is stated plainly, and no claim overstates what the method can support.
Reproducibility
An independent examiner, given the record, could follow it and reach the same result.
//Who keeps this page current
Matt Aubin, CDFE, FBCI
AI investigation expert and cybercrime specialist. Founder, Southern Recon Agency and E3 Legacy Intel. Investigating since 2009. Certified Digital Forensics Examiner, Florida Board Certified Investigator, and the builder of the proprietary AI investigation platform his own firms run on live cases. He works behind counsel as the consulting expert on AI assisted and AI generated evidence: he validates or attacks it, writes the analysis, and briefs the attorney. This page was last reviewed against the committee's materials on September 19, 2026.
AI touched the evidence in your matter?
Send the evidence and the deadline. You get a written analysis for counsel, priced on the value of the work, scoped in writing.