Deepfake Laws by State: Legislation, Acts & Sexual Content

A woman in Prague finds out her face is on a porn site. She never took that photo. She never posed for it. Somebody fed her real face into an AI tool, and now a stranger's body wears it, and it's out there for anyone to see. That's not a hypothetical — it's the exact kind of case that's led Czech police to open 35 criminal investigations in just six months, all involving AI-made intimate images shared without the person's consent.

Czech police opening 35 criminal cases in six months over deepfake porn is a real-time preview of where deepfake laws by state and by country are heading — fast, uneven, and years behind the tech that created the problem.

Here's the part that should stop you mid-scroll: 35 is not the number of victims. It's the number of cases where someone already reported it, police already opened a file, and the process is already moving. Every study on this kind of abuse says the real number of people targeted is much higher than the number who ever tell anyone. So when you read "35 cases," read it as a floor, not a ceiling.

Deepfake Laws by State and Country: Why Czechia's New Legislation Is a Preview, Not an Outlier

As of January 1, 2026, Czech criminal law added a specific offense: misusing someone's identity to make or spread pornography — including AI-made fakes, according to Dostupný advokát, which covers the new Section 191a of the Czech Criminal Code. It criminalizes producing, distributing, and circulating this material online, and it's part of a broader legislative package that also punishes sharing real intimate photos without consent — with penalties running up to five years in prison. Similar legislation is being drafted or debated in dozens of other countries right now, which is exactly why disclosure requirements and clear definitions of consent matter so much to how any new law actually functions in practice.

Think about what that timing means. The law went live January 1. Six months later, police already had 35 open cases. That's not a slow trickle — that's a dam breaking the moment there was somewhere for the water to go. People had been living with this harm for a while (artificial intelligence tools have been widely available for years now), but there wasn't a clean legal box to put the complaint in. Give people that box, and the reports show up almost immediately.

Czechia isn't alone in scrambling to build that box. Non-consensual deepfakes are now a criminal offense in more than 49 countries around the world — a number that's more than doubled since 2022, according to research covered by Expert Zoom. That's the honest state of deepfake laws by state, region, and country right now: a patchwork of state laws and national legislation that's expanding fast but still leaves huge gaps depending on where you happen to live. Anyone trying to track a specific deepfake statute state by state runs into the same problem — a law tracker that's accurate today can be outdated within months, because new bills move quickly once one state passes something. This article is part of a series — start with Deepfake Audio.

Deepfake laws by state versus national rules: what actually changed, and where a law tracker helps

In the U.S., this plays out state by state — some states have criminal penalties for non-consensual AI porn, some only offer civil lawsuits under existing state law, and some have basically nothing on the books yet. A handful of states have also taken up the political deepfake problem separately, passing disclosure requirements for AI-altered images of candidates near elections, which is a distinct issue from intimate-image deepfakes but often gets grouped into the same state deepfake conversation. Czechia took the national route: one law, one country, one court system to enforce it. That's cleaner, but it's not automatically better. A national law still needs police who know how to investigate it, prosecutors willing to bring it, and victims willing to sit through the process. Czechia just proved the first three parts can move quickly. The fourth part — victim willingness — is where things get complicated.

That withdrawal rate is the real story hiding inside the 35 cases. It's not that police can't find suspects — plenty of these cases do identify someone. It's that once a victim sees what a prosecution actually involves (giving statements, having the images entered as evidence, sometimes facing the person who did it), a lot of people decide it's not worth it. Which, honestly? Who could blame them. The system asks victims to relive the harm in order to prove it happened, and no state deepfake statute changes how exhausting that process feels from the inside.

How Common Is AI Image Harassment, and Who Gets Targeted in States With Weak Legislation?

AI image harassment is more common than most people assume, and it overwhelmingly targets women and girls — including minors, whose images are increasingly swept into the same synthetic media used for adult intimate content. Research cited by The Global Statistics puts the share of all deepfakes that are pornographic at around 98%, and the vast majority of that material is made without the subject's consent. This isn't a rare prank between bored teenagers. It's a pattern with a clear gendered target, running at scale, across platforms that mostly weren't built to catch synthetic media before it spreads.

A 2024 cross-country survey presented at the CHI Conference (a major academic conference on how people and technology interact) looked at awareness and victimization of deepfake porn across 10 countries, according to the CHI 2024 proceedings. The finding that should worry you isn't just "this happens" — it's that most people surveyed had heard of synthetic media and deepfake porn existing, but far fewer knew what to do if it happened to them, or to their kids, since minors are rarely the focus of awareness campaigns built for adults. That gap between awareness and action is exactly where victims get stuck, in states with strong laws and states with none at all.

Embarrassment or fear of further harm can make individuals reluctant to report incidents of AI-enabled abuse. — research summary cited in analysis of intimate image abuse cases, Refuge UK

AI image harassment reporting gap: why so few victims come forward under current laws

Shame is the biggest lock on the door. Most people who've had a fake made of them worry that reporting it means more people will see the image, not fewer — that filing a police report somehow makes it more real. Add in the fact that platforms are painfully slow to remove this content (more on that below), and you get a situation where staying quiet feels, wrongly, like the safer option, even in states where the laws on the books are actually fairly strong.

The Platform Problem No State Law or Legislation Is Fixing Fast Enough

Here's where it gets genuinely infuriating. An audit study covered by arXiv tested how platforms respond to two kinds of takedown requests: copyright claims and non-consensual nudity reports. Copyright claims got images removed 100% of the time within 25 hours. Non-consensual nudity reports? Zero percent removal after three weeks. Read that again. A company protecting its copyright gets same-day service. A woman reporting that her face was put on a porn image, without her consent, gets nothing — for three weeks straight, and probably longer after that. Platforms have built fast lanes for the harms that cost them money and left the slow lane for harms that cost real people their dignity, no matter what states or countries have passed on paper. Previously in this series: Biometric Login Ai Rebuilds Fingerprints From One Photo Podc.

That gap is why criminal law matters here, even when it's slow and imperfect. If the platform won't act, a criminal case at least puts the matter in front of someone with the power to compel action — a prosecutor, a court, a judge who can order removal and pursue the person who made it. In the UK, though, the picture isn't much rosier: data from Refuge UK shows that in 56% of cases where a suspect was identified, nobody was ever charged. The law existing on paper and the law actually protecting you are two very different things, whether that law comes from state legislation or a national act passed by parliament.

What You Can Actually Do Before an AI Image Spreads, Regardless of Your State's Laws

If you've ever wondered whether a photo or profile claiming to be someone is really them — a match request, a supposed "friend" sending your teenager a picture, a video that looks almost right but not quite — that worry is exactly the question this whole field of technology exists to answer. You don't need to become a digital forensics expert overnight. But there is one concrete thing worth doing now, before anything spreads: run a reverse image search on any photo of yourself or your kid that's public — a profile picture, a school photo posted online, anything a stranger could screenshot — and see where else it already appears. If a version of that photo shows up somewhere you never posted it, that's your early warning sign, and it's a lot easier to act on a warning than to undo a fire that's already spread. This matters even more for minors, since a single leaked school photo can end up fed into synthetic media tools without anyone in the household knowing until it's already circulating.

AI image harassment: what a real fake usually gets wrong

Even convincing fakes tend to have tells. Skin can look strangely airbrushed — reviewers often describe it as appearing too smooth, with none of the texture a real photo would have. Lighting on the face frequently doesn't match the lighting on the rest of the scene. Edges around hair or jewelry can blur or shimmer. None of these tells are foolproof, and the technology improves every few months, but they're worth knowing before you take an image at face value, whether the target is an adult or one of the minors increasingly caught up in this kind of content.

Thirty-five cases in six months isn't a scandal. It's a country finally counting something it used to ignore. The real number — the one nobody will ever fully know — is every person who saw their own face staring back at them from somewhere they never posted it, and decided the fight wasn't worth what it would cost them to fight it. California and a growing list of other states are watching Czechia closely, because whatever legislation comes next will need to answer the same question: how do you protect people without asking victims to relive their own worst day just to prove it happened.

Deepfake Laws by State: Frequently Asked Questions