Is That Video Real? Federal Judges Weigh a Deepfake Evidence Rule as AI Improves

More cameras, and more doubt about what they show

Almost everything people do in public is now captured on camera. Home doorbell and security systems such as Ring and Nest, business security cameras, publicly owned cameras, Flock license-plate readers, and store cameras, including those inside grocery stores like Publix, record daily life. Police body cameras add more footage, and drones are expected to bring more aerial video from law enforcement. On top of that, nearly everyone carries a phone that can record high-quality 4K video.

At the same time, artificial intelligence can now produce video and photos that are nearly indistinguishable from the real thing. In Attorney Ron Frey’s view, that combination is changing how people react to what they see. For years, many people believed whatever appeared on video. Now skepticism is rising, and it is often hard to tell whether an image or clip is real, has been altered by AI, or was fabricated entirely.

The Boston hearing: AI companies warn federal judges

On October 15, 2026, representatives from OpenAI and Meta spoke to the U.S. Judicial Conference’s Advisory Committee on Evidence Rules at a meeting in Boston. According to Investing.com, Shira Anderson, policy counsel at OpenAI, told the committee she does not believe a scientific method will exist in the future to detect AI images. Devin Norder, director of product management at Meta, said that media created by recent AI models cannot be distinguished by humans from authentic content, and that making it now takes minimal cost and time.

Watermarking is one proposed safeguard, and new laws in California and the European Union require it. But Anderson told the committee that each company uses its own system, so no universal method exists to reliably detect every watermark or verify every image or video.

So far, deepfake disputes appear to be rare in court. A March survey by the Federal Judicial Center found that only 15 of 931 responding judges had encountered a case where a party challenged evidence as a deepfake. U.S. District Judge Jesse Furman of Manhattan, who chairs the committee, noted that the issue has not yet appeared frequently in actual cases.

What proposed Rule 901(c) would do

The committee has been studying deepfakes since 2023. According to its October 2026 agenda book, it has prepared, but not finally approved, a proposed new Federal Rule of Evidence 901(c). The draft works in two steps. First, the party challenging an item as fabricated would have to present evidence sufficient to support that finding. If that showing is made, the burden would shift to the party offering the item to show that it is more likely than not authentic. The draft also calls for reasonable pretrial notice of the intent to make such a challenge, unless the court orders otherwise.

The proposal is not settled. The agenda book records that the Department of Justice’s representative said the rule is not ready for publication, and at least one committee member suggested it was premature.

Two hurdles: authentication and the weight a jury gives evidence

As Attorney Ron Frey sees it, these developments challenge courts at two levels. The first is authentication, which is foundational to introducing any evidence. Before a video or photo can be used, the party offering it generally has to show it is what they claim it is.

The second comes after authentication. Even when evidence is admitted, a question remains about how much weight a jury will give a video or photo. Jurors who know how convincing AI fakes can be may view footage differently than jurors did a few years ago. Attorney Ron Frey expects these challenges to grow as AI capability increases.

What this means for people accused of internet and computer crimes

Internet and computer crime cases often depend on digital evidence: screenshots, chat logs, images, recordings, and files pulled from phones and computers. Questions about where a file came from and whether it was altered may become more important in both Florida state courts and federal court. In practice, that can mean:

• Authentication: Metadata, device records, and the chain of custody can matter when deciding whether evidence is what it claims to be.
• Forensic review: An independent forensic examination of devices and files may help show how an item was created, edited, or shared.
• Defense challenges: If proposed Rule 901(c) is adopted, a person challenging federal evidence as AI-generated may need to come forward with supporting evidence and give pretrial notice. Florida courts follow their own evidence rules, and how they will address deepfake claims is still developing. Learn more about AI-generated images and deepfakes.

The agenda book also notes that, according to the committee’s reporter, Rule 901(c) would not cover AI-generated images offered to prove an element of a child exploitation offense when the government concedes the images are fabricated, because that would not involve a fake being offered as something it is not.

Talking to an attorney

If you are under investigation or have been charged with an internet or computer crime, the evidence against you deserves careful review. Attorney Ron Frey generally advises people to remain silent and speak with an attorney. Every situation is unique, and contacting an attorney is a priority to get tailored advice.

Attorney Ron Frey of The Frey Law Firm, LLC, a conference speaker and published author, represents people facing internet crime charges in Florida and in federal court. Contact us or call (844) 766-3739 nationally or (813) 775-2276 in Florida.