Make Deepfake Videos, Create Deepfake Images: Pa. Trooper Jailed

A fake intimate image of you can exist even if you never posed for anything. People who make deepfake videos and images need only a few ordinary photos of a face. A former Pennsylvania state trooper just went to prison for doing exactly that, and the details should make every parent and every person with a social media account sit up.

Make deepfake videos and fake images is something one man did at scale in Pennsylvania, and the sentence of 5 to 10 years shows courts now treat a fabricated image as real harm, with the person pictured never to blame.

Why anyone can now make deepfake videos from a single photo

Here is the case, in plain terms. According to the Philadelphia Inquirer, former state trooper Stephen Kamnik was sentenced to 5 to 10 years in prison. He had child sexual abuse material. He also used AI to build deepfakes (fake images or video made by software to look real). Investigators found more than 10,000 pornographic images on his work computer.

The detail that should bother you most is where some of the faces came from. Per the research on this case, he reached into PennDOT's driver's license and ID photo databases. Those are the pictures people hand over at the counter for a license. Nobody agreed to anything else.

Create deepfake videos with a face swap: what that means

A face swap (a program that pastes one person's face onto another body in a photo or video clip) used to need real skill. Now it is a tap-and-upload job in many apps. To create deepfake videos, a person only needs a face and a source clip. That is why this story is not really about one bad man. It is about how low the bar has dropped.

What an AI video generator is, in plain English

An AI video generator (a program that turns a still picture or a line of text into moving video) does the heavy lifting. A deepfake video is simply the output when that tool is aimed at a real person's face. The photo does not have to be nude. The nudity is invented. The face is the only real part, and the face is all the harm needs. This article is part of a series, start with Biometric Face Scan Brazil Ties 280m Phones To Verification.

Read that again. In a school of 1,700 teens, that is roughly 100 kids. (Yes, I did the math twice.) It is also why the old advice, "just don't share anything risky," fails here. The kid in the fake never shared anything.

---

Pennsylvania deepfake sentencing: how an ai face got built from state records

The sentencing matters because of who did it. A trooper is someone the public trusts with sensitive records. He worked on state computers, and the research describes thousands of fabricated images. The check that should have caught it, an audit trail (the record of who looked at what, and when), clearly did not stop him in time.

A real ai face built from a real license photo is a nasty combination. The face belongs to a real person with a real name and a real identity (who you are on paper and in your community). The body is fiction. The damage lands on the real person anyway.

An intimate image doesn't have to be real to cause real harm.

CaraComp commentary, on the Pennsylvania case reported by the Philadelphia Inquirer

Why an ai face swap is hard to trace

Investigators have a trick for old abuse material. They use hash matching (turning a known image into a digital fingerprint, then scanning for copies). It works when the picture already exists somewhere. A brand new fake has no earlier copy, so there is nothing to match. An ai face swap produces a fresh file every time.

Question	Older abuse images	AI deepfakes
Can hash matching find it?	Often yes, if a copy is known	Usually no, because each fake is new
Where does the face come from?	A real photo of a real victim	Any photo, even a license photo or an upload from a school page
How fast is it growing?	Slower, one image at a time	More photorealistic deepfakes in the first half of 2026 than all of 2025
Who is to blame?	The person who made or shared it	The person who made or shared it, never the person pictured

One number from the research stands out. Children aged 7 to 13 made up 79% of the AI-generated abuse material identified in the first half of 2026. That is the age group of kids who have a class photo and not much else online. Courts, meanwhile, have no single agreed standard for accepting AI detection results as evidence. A case can hinge on proof that is hard to get.

---

Can someone create deepfake videos of a real person without a script or skills?

Mostly, yes. Many tools now ask for little more than a photo and a few typed words, a script (the lines the fake person will say) or a text prompt. That does not make it legal or harmless. It means the barrier is curiosity and bad intent, not talent, which is why families are right to worry. Previously in this series: Age Verification Turkey Keeps Ids Away From Apps On Nov 1.

Where avatar tools fit, and where they do not

You may have seen avatar tools in this conversation. The reporting on the Pennsylvania case does not name any specific product. Tools that animate a photo into a talking figure are a wide category, and this story is about what one person did, not about any one brand.

What a talking avatar can and cannot prove about identity

A personal avatar (a digital stand-in for a person, often a talking cartoon character or a lifelike face) can look convincing and still prove nothing. A realistic voice or a smooth face is not evidence that the person was there. Seeing is no longer believing. That cuts both ways: real victims sometimes get doubted, and real images sometimes get dismissed as fake.

What to do if a fake intimate image of you or your child shows up

First, breathe. Then act in this order. Record the web address and date. Do not download or share sexual images of a child; ask police how to preserve evidence safely. Do not reply to the person who posted it. Use the platform's report tool to ask for removal, and contact local police, because anyone who would create this kind of material may be committing a crime. Tell one trusted adult. Do not delete your own evidence before asking how to preserve it.

If you have ever wondered whether a photo or a profile is really who it claims to be, that is the exact question facial comparison (software that measures how alike two faces are) exists to answer. For the investigators who get these cases, the research notes that the real face behind a fake can sometimes be traced back to a source photo. That is a comparison problem, and a fast one to solve with the right tools.

Here is the question that stays with me. The man who could make deepfake videos from PennDOT's photo counter was the one paid to protect those records. Who is watching the watchers when your face is the file? Up next: Make Deepfake Videos Pa Trooper Gets Prison For Face Swap.

make deepfake videos: Frequently Asked Questions