Media Literacy

Deepfake Detection Basics for Everyone

AI can now put words in anyone's mouth. Here's what to look for — even if you're not a tech expert.

Imagine a very talented impressionist who can not only mimic someone's voice perfectly, but also put on a mask that looks exactly like their face. From a distance, you'd swear it was the real person. Deepfakes are the digital version of that — and the masks are getting better every year. The good news is that skilled fakers still leave traces. You just need to know where to look.

The word "deepfake" combines "deep learning" (the AI technique involved) with "fake." The technology works by training AI on thousands of images or video clips of a real person, then generating new footage of that person in any situation the creator invents.

You don't need to become a forensic analyst. Most deepfakes that you'll encounter in everyday life — shared on social media, sent via messaging apps — still have telltale signs you can catch with the naked eye and a little practice.

Where Deepfakes Show Their Cracks

AI struggles most with the small details of real human faces, particularly at the edges and in areas that change rapidly with emotion. Here's a map of where to look:

The Deepfake Tell-Tale Map

Hard for AI
Eyes: Blinking patterns often wrong — too infrequent or unnatural. Reflections in pupils may not match. Gaze doesn't always track naturally.

Teeth and mouth: Gaps between teeth, individual tooth shape, and the way lips meet are frequently blurry or distorted.

Hair edges: Individual strands near the boundary of face and hair are often "melted" or blurry — especially flyaways.
AI Does Well
Skin texture: AI has gotten very good at realistic skin — pores, color, lighting. Don't rely on "it looks too smooth."

General face shape: Proportions are usually accurate and consistent. The overall impression of a face is convincing.

Expressions: Basic expressions (smile, neutral) are well-handled. Extreme emotions or rapid changes are harder.
Background Clues
Ear details: The complex shape of ears is often wrong or inconsistent between frames.

Jawline boundary: Where face meets neck or collar can show a faint halo, color mismatch, or blurring.

Background text: Any text visible in the scene (signs, labels, books) is often garbled nonsense on close inspection.
Motion Clues
Head movements: Abrupt or unnatural head motion, especially when turning side to side, often reveals a deepfake "seam."

Lip sync: Mouth movements may be slightly off from the audio, particularly on consonants (B, P, M).

Jewelry and glasses: Earrings, necklaces, and glasses frames often warp or flicker near the face boundary.

Real vs. Fake: A Side-by-Side Guide

Signs of a Real Video

  • Consistent lighting as the person moves
  • Natural, irregular blinking patterns
  • Sharp, detailed hair including flyaways
  • Consistent ear shape and detail
  • Lip movement closely matches speech sounds
  • Background stays consistent through motion
  • Compression artifacts are uniform across frame
  • Reported by multiple news sources if news-relevant

Warning Signs of a Deepfake

  • Lighting shifts or flickers on the face only
  • Reduced blinking or robotic blink rhythm
  • Blurry or "melted" hair at edges of face
  • Mismatched or inconsistent ear shape
  • Teeth appear blurry or abnormally shaped
  • Subtle halo or color mismatch at face edges
  • Background text is garbled or unreadable
  • Only shared on low-credibility sites

The Five-Step Reality Check for Suspicious Videos

When you see a video that seems shocking or too good to be true, run through these steps before you share or believe it.

1
Check the source

Where did this video originate? A verified news account with a long track record is very different from a three-week-old account with no history. Pause on anonymous or new accounts sharing shocking content.

2
Search for the story elsewhere

If a video shows something genuinely newsworthy — a politician confessing, a celebrity doing something scandalous — major news organizations would be all over it. A Google search for the person's name plus the claim takes 30 seconds and can save you from spreading false information.

3
Pause the video and examine the face

Watch in slow motion if possible. Focus on the jawline, hair edges, and teeth. Look at the background for any text. One suspicious detail alone may not mean much, but two or three together are a strong signal.

4
Use a reverse image search

Take a screenshot from the video and run it through Google Images (images.google.com). This often reveals whether the footage is borrowed from a legitimate older video and edited, or shows it appearing in other suspicious contexts.

5
Ask AI to help analyze it

Several AI tools can flag deepfake probability. Microsoft Video Authenticator, Deepware Scanner, and Intel's FakeCatcher are available to the public. None are perfect, but they add one more data point to your assessment.

Use AI to Help Evaluate What You See

AI assistants can be useful partners in thinking through suspicious content. Try these prompts:

"I saw a video that claims [brief description of claim]. What should I check to determine if this is genuine? What sources would be most reliable for verifying it?"
"What are the most current and reliable tools for detecting deepfake videos? Are there any that work in a browser without downloading software?"

AI assistants won't watch the video for you — but they can help you build a verification checklist, suggest fact-checking resources, and explain what detection tools exist and how they work.

Audio Deepfakes — The Less Visible Threat

While video deepfakes get the most attention, voice cloning has become just as sophisticated and is used just as often. AI can now clone someone's voice from as little as a few seconds of audio.

The key defense against audio deepfakes is establishing a verification word or phrase with people you care about — something only you two would know that you can use as a "prove it's you" check in suspicious situations.

Spies in old movies had code words — a phrase known only to allies to confirm identity. In an age of voice cloning, a simple family code word ("If you ever call me in an emergency, mention our dog's name first") can protect against very convincing scam calls.

Trusted Resources for Going Deeper

Common Questions

What exactly is a deepfake?
A deepfake is an AI-generated video, image, or audio clip that makes a real person appear to say or do something they never actually said or did. The term comes from combining "deep learning" (the AI technique used) with "fake." Early deepfakes were obvious; current ones can be remarkably convincing.
Are deepfakes always malicious?
No. Many deepfake-like technologies are used for entertainment, movie special effects, accessibility tools (giving speech to those who have lost their voice), and even education. The same technology that creates a convincing movie de-aging effect can be misused to put false words in a politician's mouth. Intent and context matter.
Can I use an app to detect deepfakes?
Several tools exist — like Microsoft Video Authenticator, Intel FakeCatcher, and Deepware Scanner — but they're imperfect and constantly catching up to new generation methods. Human visual inspection using the clues in this guide remains an important first step, and context checking (does any legitimate news source report this story?) is often more reliable than any detector.
What's the most reliable deepfake tell?
Context is the most reliable tell — not visual artifacts. If a video shows a famous person saying something shocking and no credible news organization is reporting it, that absence is more meaningful than any pixel-level detail. AI-generated images and video are improving faster than detection tools, but the context surrounding a claim changes much more slowly.
What should I do if I see a deepfake of someone I know?
Don't share it. Report it to the platform where you found it — most major platforms have specific deepfake reporting options. If it's causing real harm to someone (non-consensual intimate imagery, for example), many states now have specific laws allowing the victim to take legal action. The National Center for Missing and Exploited Children and the Cyber Civil Rights Initiative both offer resources for victims.