That AI Blog
Is this photo real? A 2-minute check before you share or pay
Your eyes aren't enough anymore
A couple of years ago you could spot a fake photo by the weird hands or the melted background. Not now. AI images and videos have got good enough that a quick look on your phone won't tell you much.
That matters when it's more than a funny meme. A photo of a ute for sale that's too cheap. A "before and after" from a contractor you've never met. A video of someone you know asking for money. A screenshot of a "breaking news" story that's doing the rounds in the group chat.
You don't need to be a tech whiz to check. You need about two minutes and a bit of healthy suspicion.
Step 1: Slow down for 10 seconds
Most scams work because they rush you. "Only one left." "Need the deposit today." "Don't tell anyone yet."
If a photo or video is pushing you to share, buy or pay right now, that's your cue to stop and check. Real deals and real emergencies can survive two minutes.
Step 2: Ask an AI tool to check it
Some AI tools can now look for hidden watermarks that get baked into AI-made images. For example, Google's Gemini app lets you upload an image or video and ask:
"Is this AI generated?"
It checks for Google's own invisible watermark (called SynthID). Here's the catch: it can only spot stuff made with Google's AI tools. If it says it can't find the watermark, that does not mean the photo is real. It might just have been made somewhere else.
So treat this as one clue, not the final answer. Check the output before you trust it.
Step 3: Search for where it came from
This is the old-school check and it still works best. Do a reverse image search:
- On your phone, open Google Lens (in the Google app) and point it at the photo or upload a screenshot.
- Or upload it to a site like TinEye.
You're looking for the earliest place the photo shows up. If the "ute for sale in Brisbane" first appeared on an overseas listing three years ago, you've got your answer. If a "news" photo only exists on random accounts and no proper news site, be careful.
Step 4: Check the source, not the picture
Ask yourself who actually posted it.
- Is it a real business with a history, reviews and an ABN you can look up?
- Is it a news outlet you've heard of, or an account that popped up last month?
- If it's a mate or family member, does the message sound like them?
Labels on social media help a bit, but they're not perfect. Platforms sometimes tag real photos as AI because of editing apps, and some fakes slip through with no label at all. Don't lean on the label alone.
Step 5: When money's involved, ring a real person
If a photo or video is part of someone asking for money, stop using the internet to check and pick up the phone. Call the person on the number you already have, not one from the message. Call the business on the number from their website.
A quick call beats any AI detector going.
The 2-minute checklist
Save this one:
- Pause. Is it rushing you to share, buy or pay?
- Ask an AI checker like Gemini, but remember it only spots its own watermark.
- Reverse image search with Google Lens or TinEye to find where it first showed up.
- Check who posted it, not just what it shows.
- Ring a real person before any money moves.
The bottom line
AI fakes aren't going away, and they'll only get better. You don't have to become paranoid about every photo. Just make the 2-minute check a habit for anything you're about to share, buy off or pay for.
If it passes all five, crack on. If it fails one, give it a miss.