Even Google themselves previously offered a no-auth way (just very niche): if you uploaded an image to Google Image search, and went to "About this image", it would show if the image was Google AI-generated. Now it doesn't.
We need to inform everyone that this technology can encode database identifiers and enough entropy to uniquely identify you as an author (or downloader).
It's presumably so they can rate-limit people who are trying to reverse-engineer or otherwise strip it (e.g. iteratively tweaking until it stops getting detected)
That’s the justification, but the EULA incorporates Google’s regular consumer terms; which means what you upload can also be used for advertising and targeting; and basically any purpose whatsoever by Google.
Yeah I've been wanting to run a ton of Wikipedia images through the detector to see if fakes snuck in. Doesn't seem to be a practical way to do this. Even Open AIs tool has a low rate limit
I've never seen it mentioned anywhere so I will just here complain that Google also removed the ability to paste images into Google Image search some years ago for no apparent reason. It worked great and I used it all the time.
It's a real shame Google isn't more transparent about how it actually works, and doesn't provide any mechanism for classifying images in bulk or offline.
I worry we’ll have to approach this the other way around: verify that photos came from a camera, using hardware support like Apple’s Reference Image, rather than try to detect every AI generated one.
In many situations, photos are evidence. AI tools make convincing fakes, such as images of defect product.
It's baffling that this requires auth. OpenAI's own tool (https://openai.com/research/verify/) doesn't require auth.
Even Google themselves previously offered a no-auth way (just very niche): if you uploaded an image to Google Image search, and went to "About this image", it would show if the image was Google AI-generated. Now it doesn't.
These are "spy"marks, not watermarks:
https://brand.io/article/spymarks/
We need to inform everyone that this technology can encode database identifiers and enough entropy to uniquely identify you as an author (or downloader).
This extends to other forms of media as well.
It's presumably so they can rate-limit people who are trying to reverse-engineer or otherwise strip it (e.g. iteratively tweaking until it stops getting detected)
The EULA you have to accept hints that they are afraid you are going to abuse it to remove synthID - e.g, set up and edit and check loop
That’s the justification, but the EULA incorporates Google’s regular consumer terms; which means what you upload can also be used for advertising and targeting; and basically any purpose whatsoever by Google.
Yeah I've been wanting to run a ton of Wikipedia images through the detector to see if fakes snuck in. Doesn't seem to be a practical way to do this. Even Open AIs tool has a low rate limit
I've never seen it mentioned anywhere so I will just here complain that Google also removed the ability to paste images into Google Image search some years ago for no apparent reason. It worked great and I used it all the time.
Click the little camera icon in the search box and it pops a modal that says Search any image with Google Lens.
The textbox below it says Paste image link, but you can actually paste an image from your clipboard here, too.
You’re a hero. So glad I complained. Why would they not just keep this working with the main text box focused…
No idea. I used it a lot too, not sure if pasting here worked from the beginning of them introducing this Google Lens thing or not.
I wonder if it’s a deliberate anti-user decision to get more URLs and less pasted images.
URLs expand Googlebot’s indexes; pasted images don’t.
this + image translate with copy+paste is the only reason I ever use yandex of all places...
I might be missing something but Google Image Search still works for me.
Any access to a watermark detector can be used to remove the watermark. Presumably they want to be able to track who is doing that.
It's a real shame Google isn't more transparent about how it actually works, and doesn't provide any mechanism for classifying images in bulk or offline.
This is the best SynthID write-up I've found so far: https://fyx.me/articles/attempting-model-extraction-of-googl...
It covers how it actually works (probably), and how to train your own classifier for it, with some seemingly decent results.
(2025)?
Why the hell do I need to sign in?
I worry we’ll have to approach this the other way around: verify that photos came from a camera, using hardware support like Apple’s Reference Image, rather than try to detect every AI generated one.
In many situations, photos are evidence. AI tools make convincing fakes, such as images of defect product.
No option to detect generated text. SynthID also watermarks text, right?
Deepwalker cracked this one long time ago
https://deepwalker.xyz/blog/evaluating-synthid-watermark-rob...
Hey OpenAi, generate a text of 100 words.
Hey Gemini, find a synonym for every second adjective. Replace in text.
Hey Grok, find a synonym for every third proper noun. Replace in text.
Hey …
is this by Google?
Yes but OpenAI, NVIDIA, Kakao, Apple have said they will also implement SynthID
This is a Google Deepmind project to “Identify AI generated media”
More detail at https://deepmind.google/models/synthid/
For those who, like me, were hoping it was a vision model to identify synthesizer models from photos of concerts and music studios!
Can’t be used without signing in with a Google, Apple, or ChatGPT account.
Also it can't detect any Synths in The Commonwealth.