To spot a deepfake, slow down and check the source before you believe or share anything. Look for telltale flaws like odd hands, blurry or warped edges, strange blinking, or audio that doesn’t quite match the mouth. Then verify: does a trusted news outlet report the same thing? Be especially careful with anything shocking or emotional, because that’s exactly what fakes are designed to be. Here’s how to protect yourself and the people you share with.
What a deepfake is, in plain terms
A deepfake is a photo, video, or audio clip created or altered by AI to show something that didn’t really happen. AI can now put words in someone’s mouth, swap faces, or clone a voice from a few seconds of audio. Some are harmless fun. Others are used for scams, misinformation, and manipulation.
The good news: you can get better at spotting them, and a few simple habits protect you from most.
Visual clues in photos and videos
No single clue is proof, but several together are a red flag:
- Hands and fingers. AI still struggles here. Look for extra fingers, odd bends, or blurry hands.
- Edges and backgrounds. Warping around the hair, ears, or where a person meets the background.
- Skin and lighting. Too smooth, waxy, or lighting that doesn’t match the scene.
- Eyes and blinking. Unnatural staring, odd blinking, or eyes that don’t track right in video.
- Mouth and audio sync. In video, watch if the lips don’t quite match the words.
- Text in the image. Garbled or nonsense text on signs or labels is a common giveaway.
The habit that matters most: check the source
Visual clues help, but fakes are getting better, so the real protection is your behavior, not your eyesight.
Before you believe or share, ask:
- Where did this come from? A random account or a trusted source?
- Is a reputable news outlet reporting the same thing? If something huge supposedly happened and only one odd post mentions it, be suspicious.
- Does it feel designed to make me angry or scared? Strong emotion is the tool fakes use to make you share before you think.
A 30-second search to confirm beats being the person who spread something false.
The scam version: cloned voices
One deepfake hits families directly. Scammers clone a loved one’s voice and call in a panic asking for money. Protect against it with one simple step: agree on a family code word. If a call claims to be a relative in trouble, ask for the code word. No code word, no money. Hang up and call the person back directly.
The common places fakes show up
Knowing where deepfakes tend to appear helps you stay alert at the right moments:
- Celebrity or politician “gotcha” clips. A famous person supposedly saying something shocking. Often fake, designed to spread fast on emotion.
- Investment and crypto scams. A trusted public figure “endorsing” a money-making scheme. Almost always a deepfake. Real opportunities don’t arrive this way.
- Fake emergency calls. The cloned-voice scam, where a “relative” calls in a panic for money.
- Romance and impersonation scams. Fake video or voice used to build trust over time, then ask for money.
- Doctored news images. Real-looking photos of events that never happened, shared to mislead or stir outrage.
The thread connecting all of these: they target your emotions, urgency, fear, excitement, outrage, because emotion makes you act before you think.
Build the pause into your family
The single best protection isn’t technical, it’s a shared family habit: pause before you believe or share. Talk about it at dinner. Show your kids and your parents how realistic fakes have become. Agree on the voice-call code word together. Make “let me check that first” a normal, even proud, thing to say. A family that pauses is a family that doesn’t get fooled, and doesn’t accidentally spread something false to everyone they know.
It’s okay to feel uneasy about this
Deepfakes are genuinely unsettling, and it’s fair to wish the world were simpler. But fear isn’t the answer, and neither is avoiding technology altogether. The answer is a few sturdy habits: check the source, watch for emotional manipulation, verify through another channel, and pause when unsure. With those in place, you can navigate this new landscape with clear eyes instead of anxiety. Awareness is the cure for fear here, and you now have it.
When in doubt, don’t share
The simplest rule of all. If you’re not sure something is real, don’t pass it on. Sharing a fake, even by accident, spreads it. Pausing is a small act that protects everyone in your circle.
Frequently asked questions
Are deepfakes illegal?
Some uses are, like fraud or certain non-consensual content, and laws are expanding. But plenty of fakes circulate freely, so you can’t rely on the law to filter them out for you.
Can tools detect deepfakes for me?
Some detection tools exist, but they’re not perfect and they lag behind the fakes. Your own habits, checking the source and pausing before sharing, are still the best defense.
Will deepfakes keep getting harder to spot?
Yes. The visual clues are getting subtler over time, which is exactly why the “check the source” habit matters more than spotting flaws.
What should I do if I find a deepfake of myself or family?
Report it to the platform, document it with screenshots, and if it’s harmful or used for fraud, contact the authorities. Don’t engage with the poster directly.
How do I teach my kids about this?
Show them how realistic fakes can be, and build the habit of checking the source before believing or sharing. We cover this in our guide on talking to teens about AI.
This is general safety guidance. For legal concerns or harmful content, contact the relevant platform and authorities.