Deepfake Laws Are Fracturing. Your Evidence May Not Survive 2026.

Twenty-six states have passed laws targeting deepfakes in elections. Not one federal law bans a deepfake political ad. And California's attempt to pass one? A court struck it down on First Amendment grounds.

That gap between state action and federal silence

That gap between state action and federal silence isn't just a policy debate. It's already changing what counts as evidence in court — and what doesn't. If you've ever been identified by a photo, verified your identity online, or even shared a video you assumed was real, this story touches you. Right now, more than a thousand A.I.-related bills are moving through all fifty state legislatures. They cover biometric data, algorithmic transparency, hiring tools, criminal justice, education. And they don't agree with each other. A facial comparison result that holds up in one state can be challenged on completely different grounds in the state next door. The question running through all of this: when the rules keep shifting beneath your feet, how does anyone — investigator, attorney, or ordinary person — trust that digital evidence is what it claims to be?

Louisiana passed a law called H.B. 178 that took effect 08-01-2025. It became the first statewide framework requiring attorneys to exercise what the statute calls reasonable diligence to verify that evidence is authentic before they can even offer it in court. That sounds straightforward, but it shifted something fundamental. Before, you could present a piece of digital evidence and let the other side challenge it. Now, in Louisiana, the person offering the evidence carries the burden of proving it's legitimate up front. For anyone who's ever been on the receiving end of a false accusation — or a misleading video — that distinction matters. It means someone has to vouch for what they're showing a judge before the judge ever sees it.

At the federal level, an advisory committee proposed a brand-new rule — Rule 707 — specifically designed to govern machine-generated evidence. The final vote on that proposal is scheduled for 05-07-2026. Even if it passes, it wouldn't take effect until December of 2027 at the earliest. So for the next year and a half, federal courts will keep relying on the existing standard — Rule 901 — which experts increasingly view as too low a bar for the age of synthetic media. That's a long window where deepfake evidence could enter courtrooms under rules that were written before the technology existed. And it's not just courtrooms. It means the viral clip you see shared a million times during election season might never face any formal test of whether it's real.

The twenty-six states that did act? Their laws don't line up. Roughly half the country now has some kind of rule about deepfakes in campaigns, but many of those laws only require disclosure — a label saying an ad was made with A.I. According to C.N.N. reporting, one state's law only kicks in thirty days before an election. Thirty days. A deepfake ad could run for months, shaping opinion, and only face legal scrutiny in the final stretch. For someone running a facial comparison or verifying a piece of digital evidence, this patchwork creates a practical nightmare. Your methodology might satisfy Louisiana's authentication requirements but fall short of what New York or California demands. And for the rest of us, it means the protections you have depend entirely on your zip code.