Deepfake vs Cheapfake: Differences and Real Threats

The deepfake vs cheapfake debate matters more than most people think. Both spread false stories and fool real viewers. But one needs costly AI, while the other needs almost nothing. This guide breaks down both in plain language. You will learn how each is made, how to spot them, and which one does the most harm today.

• What are deepfakes and cheapfakes?
• Deepfake vs cheapfake comparison table
• How each fake is made
• Real-world examples
• How to spot a cheapfake
• Is it legal to make them?
• Verdict: which is the bigger threat?
• Frequently asked questions

What are deepfakes and cheapfakes?

A deepfake is fake media made with artificial intelligence. The AI learns from real photos, video, or audio. Then it builds a new clip that looks or sounds real but is not.

A cheapfake, also called a shallowfake, is simpler. It uses basic edits instead of AI. Common tricks include slowing a video, cutting clips short, or adding a false caption.

Researchers Britt Paris and Joan Donovan coined the term "cheap fakes" in a 2019 Data & Society report. They warned that you do not need fancy AI to fool people. Simple edits work too, and often better.

The key insight most people miss

Many people only fear high-tech AI fakes. But experts say cheapfakes cause more documented harm right now. They are cheaper, faster, and far more common. So both deserve your attention.

Think of it like home security. People worry about a master thief picking a smart lock. Yet most break-ins use an unlocked window. Cheapfakes are that unlocked window.

This guide treats both fakes with equal care. You will learn the made, the cost, and the harm of each. By the end, you can judge any clip with more skill.

Deepfake vs cheapfake comparison table

This table shows the main differences at a glance. Use it as a quick reference when you study any suspicious clip. Each row covers one key way the two fakes differ.

How each fake is made

Deepfakes use AI models that study thousands of images or audio clips. The model learns a face or voice. Then it makes new content that copies that person.

This process once needed strong computers and skill. Now free apps make it easier. Still, a good deepfake takes more effort than a cheapfake.

There are a few main types of deepfakes. Each copies a different part of a person:

• Face swaps that place one face onto another body.
• Voice clones that copy how someone speaks.
• Lip-sync fakes that change the words a person says.
• Full synthetic people who never existed at all.

As the AI improves, these fakes get harder to spot. That is why deepfakes are the bigger long-term worry. But the tech still leaves clues, which we cover below.

Cheapfakes need almost nothing

Cheapfakes skip AI completely. An editor may slow a video to make speech sound slurred. They may crop a photo to hide key context.

Some cheapfakes use no editing at all. A real clip with a false caption is a cheapfake. This is why they spread so fast and wide.

Common cheapfake tricks include several simple moves:

• Slowing or speeding a video to change how speech sounds.
• Cutting a clip short to hide what came next.
• Adding a false or misleading caption.
• Showing an old, real event as if it just happened.
• Cropping a photo to remove key context.

None of these need a single line of code. That low bar is the whole point. It means almost anyone can make one in minutes.

Real-world examples

Real cases show why both fakes matter. They also show why cheapfakes often win in sheer volume. Two famous cases make the point clearly.

A famous cheapfake

In 2019, someone slowed a real video of Nancy Pelosi. The edit made her sound like she was slurring her words. It gained 2.2 million likes on Facebook in 48 hours, per PolitiFact.

Digital forensics expert Hany Farid said the trick was striking. "It is striking that such a simple manipulation can be so effective," he noted. No AI was used at all.

A famous deepfake

In 2024, a robocall used an AI voice clone of President Biden. It told New Hampshire voters to skip the primary. The FCC proposed a $6 million fine over the scheme.

This case shows the rising power of AI fakes. To learn more, see our pillar guide on deepfakes and our look at Grok deepfakes.

What these cases teach us

Both fakes targeted trust in public figures. Both spread fast on social media. But the cheapfake reached millions with zero AI cost.

The deepfake robocall was more advanced and personal. It cloned a real voice to reach voters directly. Yet it was caught and fined within months.

The lesson is clear. A simple edit can do as much damage as a complex one. Never judge a fake by how high-tech it looks.

Cheapfakes also tend to last longer online. They are easy to remake and repost again and again. One debunked clip often returns with a new caption.

Deepfakes can be harder to mass-produce at that scale. But each one can feel more personal and real. That mix of reach and realism is why both stay dangerous.

How to spot a cheapfake

You can catch most cheapfakes with simple checks. You do not need special tools. You just need to slow down and question what you see.

Use these quick steps:

1. Find the original source and the full, uncut clip.
2. Check the real date and place of the event.
3. Compare the caption to what is actually shown.
4. Look for speed changes that distort voices.
5. Search trusted fact-checkers like Reuters or AP.

For AI fakes, the signs differ. Watch for strange blinking, blurry face edges, or flat, robotic audio. Our guide on how to spot a deepfake covers these in depth.

A simple test for any clip

You can run one quick mental test on anything you see. Ask three short questions before you trust or share it.

First, who made this and why? Second, where is the original, full version? Third, does a trusted source confirm it?

If you cannot answer these, do not share the clip. This habit alone stops most fakes from spreading through you. It works for both cheapfakes and deepfakes.

Simple habits that protect you

A few habits cut your risk fast. Pause before you share any shocking clip. Strong emotions are exactly what fakes try to trigger.

Then do a quick reverse image or video search. This often reveals the real source and date. If a claim seems too perfect, treat it with doubt.

Watch for low-quality AI clutter too

Not all fake media tries to copy a real person. Some is just mass-produced junk. This flood of low-effort content is a growing problem on its own. Learn about this trend in our AI slop explainer.

Is it legal to make them?

The law treats these fakes in different ways. Much depends on intent and harm. The format alone rarely decides the case. Rules also vary widely from state to state.

Cheapfakes are rarely illegal just for existing. But they can break other laws. A cheapfake used for fraud or defamation can lead to charges.

Deepfakes face newer, stricter rules. Some U.S. states ban deepfakes in elections or non-consensual images. Federal rules are growing too. See our deepfake laws guide for the latest details.

The 2024 Biden robocall case shows how this works. The fake itself drew a large FCC fine. The penalty came from caller ID and election rules, not just the AI.

So both fakes can lead to real trouble. The harm and intent often matter more than the tool. If a fake is used to cheat or smear, the law can apply.

Verdict: which is the bigger threat?

Cheapfakes are the bigger real-world threat today. They are cheaper, far more common, and tied to more documented harm than deepfakes. Anyone can make one in minutes with no AI skill. Deepfakes are more convincing and rising fast, so they are the bigger future risk. The smart move is to fear both, and to never assume a fake must be high-tech to fool you.

In short, do not only watch for fancy AI. A slowed video or a false caption can spread just as fast. Media literacy is your best defense against both.

Why this balance matters

If you only fear deepfakes, you let cheapfakes slip by. If you only fear cheapfakes, AI fakes will catch you off guard. A balanced eye protects you from both.

This is the core takeaway from Data & Society's research. Do not be dazzled by the high-tech threat. Stay just as alert to the simple, low-tech one.

Frequently asked questions

What is the difference between a deepfake and a cheapfake?

A deepfake uses AI to create fake video, images, or audio. A cheapfake, or shallowfake, uses simple edits like slowing video, cutting clips, or false captions. Deepfakes need more tech and skill. Cheapfakes need little or none.

Are cheapfakes more common than deepfakes?

Yes. Researchers at Data & Society note that cheapfakes are cheaper and easier to make. They appear far more often online. They are also linked to more documented real-world harm than deepfakes so far.

How do you spot a cheapfake?

Find the original, full clip and check its real date and source. Compare the caption to what the video truly shows. Watch for speed changes that distort voices. Trusted fact-checkers like Reuters and AP can confirm the truth.

Is it legal to make a deepfake or cheapfake?

It depends on use and harm. Cheapfakes are rarely illegal alone but can break fraud or defamation laws. Deepfakes face newer bans, especially for elections and non-consensual images. Laws vary by state and are changing fast.

Conclusion and next step

The deepfake vs cheapfake gap is real, but both can mislead you. Cheapfakes win on volume today, while deepfakes grow more convincing each year. Stay sharp by checking sources and questioning captions. Next, read how to spot a deepfake and subscribe to our briefing for the latest on AI harms.