United States (federal courts) · court rule
FRE 707 (proposed): what it requires, and the records to keep
Proposed Rule 707 would make AI or machine-generated evidence offered without an expert meet the Rule 702 reliability requirements. It was published for comment in August 2025, drew mixed comments, and in May 2026 the Advisory Committee chose to revise and keep studying it rather than send it forward, so it has no effective date.
01 / What it requires
What FRE 707 (proposed) requires
Would apply to parties offering AI or machine-generated evidence in federal court without a sponsoring expert witness.
- 01As published: machine-generated evidence offered without an expert must satisfy Rule 702(a)-(d) if it would need Rule 702 when given by a witness.
- 02Revised draft: proponent shows the AI evidence helps the trier of fact, rests on sufficient data, uses reliable methods, reliably applied.
- 03Revised draft: ordinarily provide an expert to explain how the AI system reliably produced the evidence; other proof only in exceptional circumstances.
- 04Revised draft: give all parties reasonable notice of intent to offer AI-produced evidence.
02 / AI work
Where it touches AI work
This proposal is written for AI output, such as model analyses, predictions or reconstructions offered as evidence without a human expert. The revised draft defines AI as a machine-based system that makes predictions, recommendations or decisions for human-defined objectives.
03 / Dates
The dates that matter
Published for public comment
Comment period closed; 59 written comments, two hearings
Advisory Committee revises draft, defers it to fall 2026 study
Standing Committee told rule will not go forward yet
Advisory Committee mini-conference on Rules 707 and 901(c)
04 / Enforcement
Who enforces it
None yet; if adopted, applied by federal trial judges through admissibility rulings.
05 / Records
Records an organisation keeps
What an auditor or an inquiry usually asks to see. Your obligations depend on your role and your facts.
- 01Validation and error-rate records for the AI system
- 02Description of training and input data sufficiency
- 03Reproducible run records showing reliable application to case facts
- 04Pretrial notice served on opposing parties
Optional evidence · where Planisphere fits
Keep a record anyone can check
When your team meets a duty, Planisphere can seal the event it reports as a receipt.
A disclosure shown, a review signed, a model change approved: the event your system reports.
The event becomes a sealed receipt, dated and attributed to whoever is signed in.
Paste the receipt into Verify. No account or key.
A record supports your compliance evidence process. It is not a legal finding or a certification. Systems that call Planisphere's gate API can also hold or refuse an action before it runs.
QUESTIONS
The questions that come next
Is AI-generated evidence admissible in federal court?+
It can be, but no special rule admits it. AI output offered through an expert must meet Rule 702's reliability requirements. Output offered without an expert must still be relevant and authenticated, and the court can require proof that the system produces accurate results. Proposed Rule 707, which would apply Rule 702's standards to such output, was not sent forward in 2026. The Advisory Committee is studying a revised draft that would presume an expert is needed to show the AI produced the evidence reliably. Read the full answer
What is proposed Federal Rule of Evidence 707?+
Proposed Rule 707 would make machine-generated evidence offered without an expert witness meet the reliability standards of Rule 702(a)-(d), if that evidence would need those standards when given by a witness. It was published for comment in August 2025. After mixed comments, the Advisory Committee in May 2026 revised it to focus on 'artificial intelligence', presume an expert foundation while allowing other proof of reliability, and add a notice requirement. The revision has not been approved or republished, so the rule has no effective date. Read the full answer
When does FRE 707 take effect?+
It has no effective date. The rule was published for comment in 2025, but in May 2026 the Advisory Committee on Evidence Rules did not recommend final approval. It revised the draft and agreed the revision would need a new round of public comment if it goes forward. It planned further study with AI experts at its October 15, 2026 meeting. Any rule would still need approval by the Standing Committee, the Judicial Conference and the Supreme Court, and would take effect on a December 1 after Congress has had time to review it. Read the full answer
How do I lay a reliability foundation for an AI model's output?+
Show that the system is reliable and was used reliably in your case. In practice, document what the model is and which version ran, the data it was built and validated on, its known error rates or test results, the inputs behind the specific output, and that the run can be reproduced. Where you can, have a qualified expert explain these points; the Advisory Committee's revised Rule 707 draft would presume one is needed. Give the other side early notice and access so it can test the result. Read the full answer
Can an audit record prove my AI tool is reliable enough for court?+
Not by itself. Whether evidence is reliable enough is for the judge to decide, under Rule 702 when an expert relies on the tool and under the general relevance and authentication rules otherwise. An audit record showing how the tool was tested, its error rates, and that a specific result can be reproduced is strong supporting evidence. It works best alongside a qualified witness who can explain the method and how it was applied to the facts of the case. Read the full answer
SOURCES
Go back to the source
Administrative Office of the U.S. Courts
Administrative Office of the U.S. Courts
CHANGE LOG
What changed, and when
Rebuilt from primary sources; facts checked on this date.
Related: FRE 707 (proposed) for law AI workflows · NIST AI RMF · OMB M-25-21/22 · DoW RAI S&IP · EO 14179 · FIPS 140-3 · SHA-256 / FIPS 180-4