TL;DR
- Proposed Federal Rule of Evidence 707 (machine-generated evidence) and proposed Rule 901(c) (deepfakes) are still under study. They were not advanced for final approval at the Evidence Rules Advisory Committee’s May 2026 meeting, and the often-cited December 1, 2027 effective date is not on track.
- The public comment period on both proposals closed February 16, 2026. At its June 2026 meeting, the Standing Committee moved only an unrelated Rule 609 change forward for final approval.
- Judges are not waiting for new rules. They are already excluding AI-altered video and requiring disclosure of AI-generated exhibits under the evidence rules on the books today.
- Your authentication plan cannot sit on the shelf until 2027. The tools to challenge or defend AI-touched evidence exist right now.
There is a significant gap between legal industry headlines and the actual status of proposed rulemaking. You may have seen reports promising a new Federal Rule of Evidence for artificial intelligence with a 2027 effective date. It is critical to understand the reality, as relying on rules that are not yet in place poses a significant risk to your case strategy.
What is the status of the proposed Federal Rules of Evidence for AI?
Two proposals have been working through the Judicial Conference. Proposed Rule 707 would treat machine-generated evidence offered without a supporting expert the way Rule 702 treats expert testimony, requiring it to meet reliability standards before a court admits it. Proposed Rule 901(c) would address deepfakes by giving courts a path to demand more authentication when a party makes a credible claim that evidence was fabricated by AI.
Both remain proposals. The Advisory Committee on Evidence Rules kept Rule 707 and Rule 901(c) as study items at its May 2026 meeting rather than sending them up for final approval. The comment period closed on February 16, 2026, and the responses are still under review. At the Standing Committee’s June 2026 meeting, the only evidence amendment advanced for final approval was an unrelated change to Rule 609.
What that means for you: no AI-specific evidence rule is in force, and the timeline often quoted on conference panels is not locked. Build your strategy around the rules that govern today.
How do courts evaluate AI-generated evidence under current rules?
Here is the part that gets lost. Courts have been handling AI-generated and AI-altered evidence for two years using existing authority, and they have shown a willingness to exclude it.
In State v. Puloka, a Washington court rejected AI-altered video, finding the method behind the alteration had not won acceptance among experts and risked misleading the jury. A New York court went further in Matter of Weber, holding that counsel has an affirmative duty to disclose when an exhibit was generated by AI, and that such evidence can trigger a reliability hearing. And in Kohls v. Ellison, a federal court in Minnesota struck expert testimony that leaned on a declaration containing AI-fabricated citations.
The common thread runs through the rules you already use. Rule 901 authentication and Rule 702 reliability, with Rule 403 there to weigh prejudice, give a judge what is needed to keep unreliable AI evidence out. The proposed rules would sharpen the focus. The ones already on the books are sharp enough to cut.
“The proposed rules would sharpen the focus on AI evidence. The rules already on the books are sharp enough to cut.”
Authentication Best Practices for AI-Generated and AI-Altered Evidence
You will meet AI-touched evidence from two directions: exhibits you want to admit, and exhibits the other side offers that may have been altered. Both call for the same groundwork.
- Authenticate your own digital evidence early. Capture and document the chain of custody for video, images, and audio at collection, so you can establish the source without scrambling at trial. This is where forensic discipline earns its keep, the same discipline behind Why Hiring a CCE-Certified Computer Examiner Matters.
- Question the source of what you receive. When a key exhibit looks too clean or too convenient, ask how it was created and captured. A credible authenticity challenge can force the other side to prove what they are offering.
- Disclose your own AI use. Matter of Weber signals that courts expect candor about AI-generated demonstratives. If you build a visual with AI assistance, be ready to explain how, a point worth weighing alongside our look at trial graphics and the role of visuals.
- Treat metadata as evidence about evidence. The data attached to a file often reveals whether it was edited, and when. Preserving it can decide an authenticity fight.
This connects to the larger shift we flagged after Legalweek, in 3 Things Legalweek 2026 Made Clear About Defensible eDiscovery: the conversation has turned from whether to adopt AI toward whether your use of it holds up under challenge. Evidence is where that question gets answered in front of a judge.
Expert Forensic Support for AI Evidence Authentication
When authenticity is contested, AI-generated evidence authentication comes down to the record you built at collection. Our forensic collections and investigations team captures digital evidence with documented methodology and preserves the metadata that proves whether a file was altered, so provenance becomes something you can demonstrate with documentation when the other side questions it. When you are on the receiving end of suspect evidence, our examiners can analyze it for signs of manipulation and give you the technical footing for a 901 or 702 challenge.
The rules will catch up to AI in time. Your matters will not wait for them. Working with a partner who treats authentication as a discipline from the first collection means you are ready for the evidence fight whenever it arrives, under the rules as they are written today.
Facing evidence that may have been touched by AI? Reach Complete Legal at info@completelegal.us or 888-885-9909 to build a forensic record that holds up when authenticity is on the line.
Contact Us:
info@completelegal.us | 
888-885-9909
Key Takeaways
- Proposed FRE 707 and Rule 901(c) are still under study; they were not advanced for final approval in 2026, and no AI-specific evidence rule is in force.
- Courts are already excluding AI-altered evidence and requiring disclosure of AI-generated exhibits under Rules 901, 702, and 403. See Puloka, Weber, and Kohls v. Ellison.
- Authenticate your own digital evidence at collection, question the provenance of what you receive, and preserve metadata that can settle an authenticity dispute.
- A documented forensic record is your strongest position in any AI-evidence challenge, today and after the rules are adopted.
