How to Spot a Deepfake: A Plain Guide for Older Adults

A deepfake is a photo or video made or changed by AI so it looks like a real person said or did something they never did. If you want to know how to spot a deepfake, here is the short answer. Look for small visual glitches, ask who posted it and why, and check the story somewhere you already trust before you react. The single best habit is the slowest one: pause before you share, send money or get upset.

I'm Andy, and I write these guides at AI Answered. Noah and I co-wrote AI for Boomers, a plain guide to using AI safely. You don't need special software. You need a few habits and a little patience.

What a deepfake looks like today

A few years ago, fake videos were easy to catch. Faces looked waxy. Lips moved out of step with the words. Those days are mostly over. Free and cheap AI tools can now make a convincing picture of almost anyone in seconds. If you're curious how easy it has become, we walked through it in our post on AI video and images.

Here are the places older adults tend to run into them:

• Family. A photo of a grandchild in a hospital bed, or a short video of a relative asking for help. These usually come with a request for money.
• Celebrities. A famous TV doctor selling a supplement, or a well-known businessman promoting an investment. The face and voice look right. The product is a scam.
• News. A dramatic image of a disaster, a politician saying something shocking, or a "leaked" clip that spreads fast on Facebook.

The pictures are getting better every month. That is why the checks below lean more on the story around the image than on the image itself.

How to spot a deepfake with your eyes

Start with the picture or video. The FBI's public warning on AI fraud lists the same signs I tell my own family to look for. Zoom in and take your time.

• Hands and fingers. Count them. Look for extra fingers, fingers that melt together, or hands bent in odd ways.
• Teeth, eyes and ears. Teeth may look like one white strip. Eyes may not point the same way. Earrings may not match.
• Edges. Watch where hair meets the background, or where a face meets a neck. Blurry or shimmering edges are a clue.
• Shadows and light. Does the light on the face match the light in the room? Do shadows fall the right way?
• Text in the background. Signs, name tags and labels often turn into gibberish in AI images.
• Movement in video. Blinking that looks off, a head that stays oddly still, or lips that don't quite match the words.
• Too perfect. Skin with no pores, a photo that looks like a magazine cover when it should look like a phone snapshot.

One caution. A clean-looking image proves nothing. Many good fakes pass every visual test. Glitches can tell you something is fake, but the lack of glitches can't tell you it's real.

Check the story, not just the picture

This is where you'll catch most fakes. Ask yourself a few plain questions.

• Who sent it? A stranger, a new account, or a friend whose account may have been taken over?
• What does it want me to do? Click a link, buy something, send money, or share it fast? Real news rarely needs you to act this second.
• How does it make me feel? Scared, angry or thrilled is exactly what fakers aim for. Strong feelings make us skip the checking step.
• Is anyone else reporting it? If a big event really happened, major news outlets will have it. Search the headline. If only one odd website has the story, be suspicious.
• Does it fit the person? Would your grandson really send a blurry photo from a hospital and ask for gift cards? Would that famous doctor really sell pills in a Facebook ad?

The National Council on Aging has a helpful guide to deepfakes for older adults that makes the same point. The context around a fake usually gives it away before the pixels do.

What to do when a family photo or video seems off

This is the scary one, so let's take it slowly.

Here's a made-up example. You get a text with a photo of your granddaughter with a bandaged arm. The message says she was in a car accident, her phone is broken, and she needs $2,000 for a lawyer right away. Please don't tell her parents.

Step 1: Don't reply to the message

Don't answer the text, click any link or call a number it gives you. That just keeps you inside the scammer's story.

Step 2: Call her yourself

Use the number you already have saved. If she doesn't pick up, call her parents or another relative. "Don't tell anyone" is a classic sign of a scam. Real family emergencies welcome more help.

Step 3: Use your family code word

The FBI suggests every family agree on a secret word or phrase. Anyone asking for urgent help should be able to say it. If you haven't set one up, do it this week at the next family phone call.

Step 4: Never pay in gift cards, wire transfers or crypto

No real lawyer, hospital or police department asks for payment that way. That one rule stops a lot of fraud.

If the same story comes as a phone call in your loved one's voice, that is a voice clone. Our post on AI scam calls covers it.

If you run a small business, the same call-back rule works for your team when a "boss" sends an urgent payment request, and our AI safety basics for business owners post explains how.

Celebrity ads and fake news clips

Celebrity deepfakes are mostly about selling you something. A familiar face makes the product feel safe.

• If a famous person is pitching a cure, a miracle supplement or a sure-thing investment, assume it's fake until proven otherwise.
• Go to the celebrity's official website or verified account. If the endorsement isn't there, it isn't real.
• Don't buy from a link inside the ad. If the product is real, you can find it on your own.

For news clips, the habit is simple. Before you share, search the headline. If it's real, you'll find it from several sources you know. If you can't find it anywhere trusted, don't pass it along. Sharing a fake, even by accident, helps it reach your friends.

Simple habits that keep you safe

You don't have to become an expert. These habits do most of the work, and they're the same rules our book uses in its section on staying safe online, from scam emails to suspicious links.

• Ask plainly. If something seems off, ask the person directly through a way you already trust.
• Check what matters. Money, health and family emergencies always get a second check.
• Keep private info private. Don't send bank details, passwords or Medicare numbers because of a photo or video.
• Slow down when someone rushes you. Pressure is the scammer's best tool. Taking an hour to check will never hurt a real emergency.

Talk about deepfakes with family before anything happens. Agree on the code word, and ask them to call you if they ever see something strange using your name or face.

Common questions

Can I use an app to tell if a photo is a deepfake?

Some detection tools exist, but none are reliable enough to trust on their own. They miss good fakes and sometimes flag real photos. Checking the story with a phone call or a trusted news source is still the better test.

Are AI photos always scams?

No. Plenty of people use AI for birthday cards and art. The problem is images made to fool you.

What should I do if I already sent money?

Call your bank or card company right away and tell them it was fraud. Then report it to the FBI at ic3.gov. Acting fast gives you the best chance of stopping the payment.

Keep learning at your own pace

Deepfakes will keep improving, but the questions that catch them stay the same. Who sent it, what does it want, and can I check it another way?

We wrote AI for Boomers so you can use these tools safely and on your own terms. It walks you through staying safe online, video calls with family and daily reminders, one step at a time, at your own pace. You can also find it on Amazon.