Google's new tool lets you spot AI-made images, sort of
Google launched a free detector for AI content in October 2026, but it only catches images made by a handful of companies.
Google rolled out SynthID Detector in October 2026, a free website where you can upload an image, video, or audio file to check whether it was made by artificial intelligence. The tool scans for invisible watermarks that Google and several partner companies embed in their AI creations. For a small business owner worried about fake content circulating online, this is worth knowing about—but understand its limits before you rely on it.
Key facts
- Google introduced SynthID in 2023 as a watermark embedded in AI-generated content, but users had no way to check for it until now.
- SynthID Detector launched for researchers and journalists in 2025 and became available to everyone in October 2026.
- The tool detects watermarks from Google's Gemini, OpenAI, NVIDIA, Kakao, and Apple (though Apple support is still in development).
- It cannot detect AI content from Meta or other AI tools that do not use the SynthID watermark system.
- Users are limited to about 10 checks per day and must log in with a Google, OpenAI, or Apple account.
What the detector actually does
The tool is simple: you go to SynthID.com, upload a file, and get back a yes or no answer. It works by looking for a specific watermark, not by analyzing the image itself for telltale signs of AI creation. This straightforward approach means fewer false alarms—it won't mistakenly flag a real photo as fake. But that same simplicity is its biggest weakness.
What it cannot do
The detector only works on content made by companies that use the SynthID watermark system. If someone created an image with Meta's tools, Midjourney, Stable Diffusion, or any of the many other AI services out there, this detector will not catch it. That covers a huge portion of the AI content actually circulating online. The tool also cannot tell you which parts of an image were AI-generated if the whole thing is a mix of real and artificial elements.
Should you use it for your business
If you're a small business owner concerned about deepfakes or AI-generated content being used to impersonate you or your company, this tool has real value—but only as one piece of the puzzle. Use it when you suspect a specific image might be fake, especially if you think it came from a major AI service. But do not assume that a negative result means the content is definitely real. Many AI tools are not covered by this detector, so absence of a watermark proves nothing.
The bigger picture: no single detector catches all AI content. Researchers have built better tools, but they are not available to the public. If you need to verify whether content is real, combine this detector with your own judgment, reverse image searches, and direct contact with the person or company involved.
The practical takeaway
For most small business owners, this detector is a useful free option to have available when you need it. It will help you confirm a suspicion in certain situations, but the technology landscape is moving faster than detection tools can keep up with. Stay skeptical of anything that seems off, especially if it involves your business or reputation. The gap between what AI can create and what detectors can catch will likely persist for some time.
Sources
- How to use Google's SynthID AI detection website - Engadget, October 18, 2026
They did the reporting. The writing and the opinions here are ours.
Questions people ask about this
Can I use this to check if a photo of me online is real or AI-generated
Yes, you can upload it to SynthID.com and check. But the result is only reliable if the image was made by Google, OpenAI, NVIDIA, Kakao, or Apple. If it was created with other AI tools, the detector will not catch it.
What AI companies does SynthID Detector work with
It detects watermarks from Google's Gemini, OpenAI, NVIDIA, Kakao, and Apple (Apple support is still being added). It does not work with Meta, Midjourney, Stable Diffusion, or most other AI services.
How many times can I check images in a day
You are limited to about 10 checks per day. You also need to be logged in with a Google, OpenAI, or Apple account to use the tool.
If the detector says an image is not AI-generated, is it definitely real
No. The detector only finds watermarks from specific companies. If an image was made with other AI tools or edited after creation, the detector may not catch it.
Are there other AI detection tools I should know about
OpenAI and Meta each offer their own detectors, but they also have limitations. No single tool reliably catches all AI-generated content across all platforms.
Should my business use this to verify customer content
It can be one helpful step, but do not rely on it alone. Combine it with reverse image searches, direct contact with the source, and your own judgment.
This piece is about How to use Google's SynthID AI detection website, published by Engadget. Go read the original; the opinions here are ours and the reporting is theirs.
