← Blog · 2026-10-18 · 6 min read

Is this video AI? A 2-minute check to share with family

What it does for you: the next time a too-cute animal clip, a shocking "news" video, or a strange celebrity quote shows up in your feed or the family group chat, you'll have a short routine to check it before you believe it or share it. And if someone you love already shared a fake, you'll have a kind way to bring it up.

This week a lot of people on X were asking some version of the same thing: "How do I get my parents to stop falling for AI videos on Facebook?" One person said both parents, nearly 60, insist every AI video they watch is real. On Reddit, someone wrote that their mom takes it personally and gets flustered every time she finds out she was fooled. That's the real problem. It isn't intelligence. These videos are made to look real and to make you feel something fast.

Here's the routine. No single step is proof, so you're looking for several signs pointing the same way.

Step 1: Notice the feeling (5 seconds)

If a video makes you go "aww," "no way," or "that's outrageous," that's your cue to slow down. Fakes are built to get a fast reaction and a fast share. Security-camera or body-cam style clips deserve extra doubt, because AI video tools are good at grainy footage like that (Axios).

Step 2: Look for a label (20 seconds)

• Facebook and Instagram: some AI posts carry an "AI info" or "AI content" label. If it's not on the post itself, tap the three-dot menu on the post. Meta puts the details for AI-edited content in that menu, not under the name (TechCrunch).
• YouTube: open the video description and look for "How this content was made." A "Made with AI" note there means it was meaningfully altered or generated with AI (YouTube Help).

No label doesn't mean it's real. Plenty of fakes aren't labeled.

Step 3: Check who posted it (30 seconds)

Tap the page or account name. Warning signs: a brand-new page, a vague name, and a feed full of the same kind of "amazing" clips with no real people in the comments. A real major event will be covered by more than one news outlet you've heard of.

Step 4: Ask Gemini to scan it (1 minute, free)

The free Gemini app has a built-in check. Sign in, upload the image or clip (videos under 90 seconds and 100 MB or less), and ask (Gemini Apps Help):

What the answer means:

• Watermark found: all or part of it was made or edited with Google's AI.
• Nothing found: it wasn't made with Google's AI. It could still come from a different AI tool. This is the most important limit to remember.

Google has also been bringing the same "Is this made with AI?" question to Google Search and Lens (Google). If you took a screenshot, crop it tight around the picture first.

Step 5: Check the claim, not the pixels

AI chatbots aren't reliable at eyeballing whether a video is fake. They're much better at searching whether the event really happened. Paste this into ChatGPT, Gemini, or Claude with web search on:

Then open the links yourself. If nobody credible covered something that big, that tells you a lot. For more practice, try the free drill Fact-check a claim with AI (safely).

Step 6: If it asks for money, stop

If a video, call, or message pushes you to send money, buy gift cards, or act right now, treat it as a scam until proven otherwise. The FTC warns that scammers can clone a loved one's voice from a short clip. Hang up and call the person back on a number you already know, or ask a question only they could answer (FTC).

Helping a parent without making them feel dumb

What works better than "Mom, that's fake":

• Say it fools everyone. Because it does, including people who work in tech.
• Make a "send it to me first" deal. If they're not sure, they forward it to you before sharing. No lecture, ever.
• Practice on their phone, not yours. Show them where the label and the three-dot menu are, then let them find the next one themselves.
• Agree on a family safe word for phone calls about emergencies.

Not sure how to say it? Let AI help you write the text:

Read it before you send it, and change anything that doesn't sound like you.

Honest limits

• No tool catches everything. Gemini's scan only recognizes Google's own watermark. Labels on Facebook, Instagram, and YouTube depend on the creator, the file's data, or the platform's detection, so many fakes slip through.
• The visual tells are fading. Weird hands and garbled text still show up sometimes, but newer AI videos often have none. Checking the source and the claim matters more than staring at pixels.
• Mind what you upload. Google says not to put anything into Gemini that you wouldn't want a reviewer to see or Google to use to improve its services (Gemini Apps Privacy Hub). Public viral clips are fine. Private family videos, skip.
• Free vs paid: everything here works on free plans. The Gemini check needs you to be signed in, and the Content Credentials part of the check isn't on the iPhone app yet, according to Google's help page.

Your 2-minute challenge: find one viral video in your feed today and run Steps 2 to 4 on it. Then show the routine to one person in your family.

Want ready-made prompts for moments like this? Free drills are on the Drills page, and the Life Scripts Starter pack ($9 founder price) includes scripts for planning a hard conversation and checking whether an AI answer is trustworthy.

Sources

• Gemini Apps Help: Verify AI-generated images, videos, and audio (SynthID and Content Credentials, file limits, what results mean)
• Google: Making it easier to understand how content was created and edited (May 19, 2026)
• YouTube Help: Understanding "How this content was made" disclosures
• TechCrunch: Meta is making its AI info label less visible on content edited by AI tools
• FTC: Scammers use fake emergencies to steal your money (voice cloning, call back on a known number)
• Axios: How to talk to seniors about AI scams
• Gemini Apps Help: Gemini Apps Privacy Hub