You know the photo.
It lands in the family group chat with three exclamation marks. A bridge collapsing. A celebrity wedding. An elephant carrying two children out of a flood.
That last one isn't made up. In August, during flash floods in Nepal, three AI videos spread online as real footage. A bridge giving way. A settlement destroyed. That elephant.
Fact-checkers ran one of them through Gemini. It found Google's hidden AI watermark.
Here's the news: since Wednesday, Google has opened that same check to everyone. You don't need Gemini, and you don't need to be a fact-checker.
It's three steps. And there's one result that fools almost everyone, which I'll get to. Right after this: the three-step check, before you hit forward.
Now a quick word from today's sponsor.
Turn AI into Your Income Engine
Ready to transform artificial intelligence from a buzzword into your personal revenue generator
HubSpot’s groundbreaking guide "200+ AI-Powered Income Ideas" is your gateway to financial innovation in the digital age.
Inside you'll discover:
A curated collection of 200+ profitable opportunities spanning content creation, e-commerce, gaming, and emerging digital markets—each vetted for real-world potential
Step-by-step implementation guides designed for beginners, making AI accessible regardless of your technical background
Cutting-edge strategies aligned with current market trends, ensuring your ventures stay ahead of the curve
Download your guide today and unlock a future where artificial intelligence powers your success. Your next income stream is waiting.
Right. Back to it.
Three steps before you hit forward
Google calls it SynthID Detector. Search for "SynthID Detector" and you'll find Google's own page.
Sign in.
Upload the photo, video or audio clip.
Read the answer.
You get a plain yes or no, and if it's a yes, which company's AI made or edited it.
Already in the Gemini app? Even easier. Upload the image and ask: "Was this created or modified using AI?"
What it can spot (and what it can't)
Content made with tools from Google, OpenAI, NVIDIA and Kakao. Apple is due to join soon.
That's because those companies stamp an invisible watermark into what their AI makes. Google says it has stamped more than 180 billion images and videos so far.
The watermark is tough, too. Google says it survives cropping, filters and compression. So someone trimming the edges doesn't hide it.
But Google is upfront that it "isn't a general AI detector." It only finds that one watermark. Which brings us to the answer most people read wrong.
The result that fools people
"Not detected."
It feels like a green light. It isn't.
Google says it plainly: "A negative result doesn't mean a file is real."
"Not detected" can mean three things:
It's real.
It was made by an AI that doesn't use the watermark. Meta, Microsoft and the free AI models anyone can download aren't covered.
Someone stripped the watermark. Researchers have already shown it can be done.
So a "yes" is strong evidence. A "no" just means keep checking. (If you remember one line from today, make it that one.)
Two free checks for when it says "not detected"
Both are free, and both are already in your phone.
Reverse image search. In Google Lens, search with the image. If it first appeared on an AI art account last week, you have your answer.
Look for the source. Real disasters get filmed by lots of people from lots of angles. One perfect clip that nobody else captured is a red flag.
(Also: elephants are heroic. But if the children are perfectly dry, be suspicious.)
Make it stick
Save this to your notes, or pin it in the family chat:
Before I forward it:
1. SynthID check (or ask Gemini: "Was this created or modified using AI?")
2. If "not detected": reverse image search in Google Lens
3. Still unsure? Don't forward it. Ask "where's this from?" insteadThat last line is the kindest one. Nobody feels silly being asked where something came from.
A quick check before you act on something. That's the habit behind The One-Person Company Playbook too: AI does the legwork, you make the call.
It's the complete setup for running a business where AI does the work of a full team. Inside:
A company folder with your brief, pipeline tracker and invoice log already laid out, so any AI picks up where you left off
Thirty AI roles across six departments, from sales to finance, ready to use
Six jobs that run on a clock: the morning briefing, follow-ups, invoice checks and more
The step-by-step path from picking a business model to your first paying customer
Works on Claude, ChatGPT, Gemini or Grok, whichever you already pay for.
The catch: it takes about eight hours over a weekend to set up. Good thing it's the weekend. $97, one-time (down from $297).
Talk tomorrow,
Zephyr
P.S. Run the last viral photo you got through it. Hit reply with what it said. I'm curious how many come back "not detected."