AI Explained

Real or AI? How to Check an Image with Gemini's SynthID (and the Catch)

You can check whether an image was made with Google's AI by asking Gemini. Here's how SynthID works, its limits, and why "no watermark" doesn't mean real.

Watch the Short

The 60-second version is on YouTube. This article is the full story.

New Shorts every week on tech trivia, AI and gadgets.

Subscribe to TheTechPrana

Key takeaways

  • Upload an image to the Gemini app or gemini.google.com and ask "Is this AI-generated?" – Gemini checks for SynthID, Google's invisible watermark.
  • The catch: Gemini can currently only recognise content made with Google's AI tools. "No watermark found" does not mean an image is real.
  • Watermarking is spreading: Google says OpenAI and ElevenLabs are bringing SynthID to more of their content, and over 100 billion images and videos already carry it.

AI-generated images are now good enough to fool most people at a glance. So when a surprising photo lands in your feed or a family WhatsApp group, how do you check it? Google offers one free way – but it comes with an important catch that most people miss.

How to check an image in Gemini (3 steps)

  1. Open Gemini – the Gemini app on your phone, or gemini.google.com in a browser.
  2. Upload the image you want to check.
  3. Ask: “Is this AI-generated?”

Gemini then looks for SynthID and tells you whether it found it.

What is SynthID?

SynthID is an invisible watermark that Google adds to images, video and audio created with its own AI tools. You can’t see or hear it – it’s woven into the content itself rather than stamped on top like a logo.

It’s also designed to be robust: it’s built to survive common edits such as cropping and resizing. Very small edits to a real image may produce an inconclusive answer, so treat borderline results with care.

What Gemini can check

According to Google’s help page, the check works on:

  • Images
  • Videos under 90 seconds
  • Audio under 1 hour
  • Files up to 100 MB

There’s also a usage limit of roughly 10 image checks a day.

The catch: “no watermark” doesn’t mean real

This is the part that matters most. Google’s help page says Gemini can currently only recognise content created with Google’s AI tools.

So if Gemini says it found no SynthID watermark, that only means the image probably wasn’t made with Google AI. It could still have come from a different AI tool that doesn’t use SynthID. In other words:

Gemini says What it means
SynthID found Made or edited with Google AI
No SynthID found Not made with Google AI – could still be AI from another tool

Never treat “no watermark” as proof that a picture is genuine.

Watermarking is spreading

This is starting to change. In May 2026, Google said companies including OpenAI and ElevenLabs are bringing SynthID technology to more of their AI-generated content, and that more than 100 billion images and videos already carry the watermark. Google is also adding this kind of AI detection to Search and Chrome.

Whether Gemini’s checker will recognise those partners’ watermarks hasn’t been confirmed yet, so for now, assume it only covers Google’s own tools.

Pro tips for better results

  • Screenshots: crop tightly around the image itself, without the app interface around it.
  • Don’t upload collages – check one image at a time.
  • Combine methods: look for the original source with a reverse image search, check who first posted it, and look for odd details such as hands, text and reflections.

The bottom line

A SynthID check is a quick, free first step that can confirm an image came from Google’s AI. It can’t prove an image is real. For anything important, check the source too – and be suspicious of images designed to make you angry or shocked.

Your turn: have you ever been fooled by an AI image? Watch the Short above and tell us in the YouTube comments.

Sources

  1. Gemini Apps Help: Verify Google AI-generated images and videos with SynthID
  2. Google Blog (20 Nov 2025): How we're bringing AI image verification to the Gemini app
  3. Google Blog (19 May 2026): Making it easier to understand how content was created and edited
  4. 9to5Google (19 May 2026): Google is adding AI detection for photos, videos and audio to Search and Chrome
AI imagesGeminiSynthIDDeepfakesHow-to