Yet another Hacker News

SynthID Detector

51 points ilreb 1 hour ago 53 comments

latexr 1 hour ago | parent

Can’t be used without signing in with a Google, Apple, or ChatGPT account.

giancarlostoro 1 hour ago | parent

Also it can't detect any Synths in The Commonwealth.

nonethewiser 32 minutes ago | parent

Who thought it was a good idea to Institute it like this?

Tiberium 1 hour ago | parent

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.

lelandfe 1 hour ago | parent

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.

RugnirViking 1 hour ago | parent

this + image translate with copy+paste is the only reason I ever use yandex of all places...

bossyTeacher 1 hour ago | parent

I might be missing something but Google Image Search still works for me.

dannyw 53 minutes ago | parent

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.

mh- 52 minutes ago | parent

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.

lelandfe 51 minutes ago | parent

You’re a hero. So glad I complained. Why would they not just keep this working with the main text box focused…

mh- 48 minutes ago | parent

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.

_flux 31 minutes ago | parent

In similar fashion in the Android Play Store they moved the search off from the top bar. For some time they educated users that "oo the search is no longer here, you need to click there -> to get to the page with search!". That has stopped (for me?), but they still haven't found any other use for same place in the first view.

I suppose it was too convenient.

apefulsin 1 hour ago | parent

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.

brokensegue 1 hour ago | parent

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

wodenokoto 1 hour ago | parent

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

dannyw 49 minutes ago | parent

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.

possibilistic 1 hour ago | parent

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.

cubefox 45 minutes ago | parent

They are not spymarks if they don't encode any personal IDs and are merely used to indicate that an image was AI-generated. I don't think Google or OpenAI use SynthID to include personal data.

morkalork 40 minutes ago | parent

And there lies the rub. You can't verify if they do or don't embed such content. These companies are just like "trust me bro".

orng 36 minutes ago | parent

Indeed and Google has built its empire through tracking users across the internet so it is really hard to trust that they won't..

MisterMunchkin 30 minutes ago | parent

Come on man, they're both advertising companies. Would you resist the opportunity to build that social graph and detect a known user anywhere on the internet?

Retr0id 58 minutes ago | parent

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)

user43928 40 minutes ago | parent

I wonder how relevant this really is.

How hard can it be to download a set of real images and generated ones in order to train a model to detect and strip the watermark with minimal perceptual difference?

cubefox 37 minutes ago | parent

OpenAI's tool is worse though because it doesn't detect SynthID from non-OpenAI models. I just tried it with various images created by Imagen / Nano Banana.

MisterMunchkin 35 minutes ago | parent

It's deliberate so that you can't find a way to bypass it.

98codes 12 minutes ago | parent

It's worth nothing that this IS Google.

Given that it's an ad company, they want something to connect what you're doing to the rest of their data, and they couldn't possibly keep the old image search results there if they want people to use this.

pbronez 1 hour ago | parent

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!

jakozaur 1 hour ago | parent

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.

sha-3 1 hour ago | parent

No option to detect generated text. SynthID also watermarks text, right?

nonethewiser 43 minutes ago | parent

Thats what I was expecting. I thought so.

tosh 1 hour ago | parent

is this by Google?

spuz 1 hour ago | parent

Yes but OpenAI, NVIDIA, Kakao, Apple have said they will also implement SynthID

nialv7 1 hour ago | parent

Why the hell do I need to sign in?

wildzzz 31 minutes ago | parent

Probably to help prevent people from reverse engineering synthID to build a filter.

Retr0id 53 minutes ago | parent

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.

MisterMunchkin 45 minutes ago | parent

They don't want you to know, because they're the baddies. Imagine how much money they can get from advertisers if they're able to identify every single piece of code, reddit thread, email, GitHub readme that you've ever written based on your secret ID. They're creaming themselves just thinking about it.

"Excellent work acquiring that outlook data Anat, now let's cross check the defunct company emails against YouTube videos edited with Google PrivateEditAI™ to figure out what the social security number of this YouTube account is."

Avicebron 40 minutes ago | parent

Unfortunately this is the truth, no one can be given a benefit of the doubt because despite years of good grace they have been anti-user and enshittified everything they touch.

elevation 44 minutes ago | parent

> It's a real shame Google isn't more transparent about how it actually works

> classifying images in bulk or offline

You've described exactly the elements spammers and fraudsters need to be able to defeat this mechanism.

Retr0id 42 minutes ago | parent

Well it's not a very good mechanism then, is it.

_flux 28 minutes ago | parent

Is it possible to implement a very good mechanism?

anon48293 52 minutes ago | parent

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 …

MisterMunchkin 43 minutes ago | parent

Ah but that's the genius of the scheme. Every single one of those providers will detect your secret ID and refuse the request. And sneak their own one in for good measure.

Muromec 36 minutes ago | parent

What's what you use the stupid silly local model^W script.

nonethewiser 8 minutes ago | parent

>Hey OpenAi, generate a text of 100 words.

Too short for watermarking but point taken of course.

Probably more like

    response = frontier_synthid_model(prompt)
    
    while frontier_synthid_validator(response) == false
        response = cheapo_chinese_model.prompt(
            "pls remove synthid",
            response
        )

m00dy 52 minutes ago | parent

Deepwalker cracked this one long time ago

https://deepwalker.xyz/blog/evaluating-synthid-watermark-rob...

ChrisArchitect 46 minutes ago | parent

(2025)? https://blog.google/innovation-and-ai/products/google-synthi...

Or did it just become public

Edit:

Blog post today https://blog.google/innovation-and-ai/models-and-research/go...

sarkarghya 46 minutes ago | parent

Here is openai:

https://help.openai.com/en/articles/8912793-provenance-signa...

There is rate limit though

cubefox 44 minutes ago | parent

There is a rate limit of around 10 checks in 24 hours.

jdranczewski 41 minutes ago | parent

Finally, the previous implementations were very silly! OpenAI had a nice tool, but it only detected their own watermarks, and Google's process was "upload the image to Gemini and ask if it's AI generated", which seemed like a perplexing waste of tokens and breath.

(While a sibling comment points out putting the image through Google Image Search as an alternative, I don't remember this being signposted in the support article I've read, so I unfortunately didn't know about it)

Computer0 5 minutes ago | parent

It also was the exact opposite of the commonly shared sentiment “Don’t blindly trust the ai”

RobGR 37 minutes ago | parent

If I go directly to the URL it wants me to agree to some stuff before I even know what the service is. I had to come to these comments to figure that out (without agreeing to anything first).

iamleppert 37 minutes ago | parent

Wouldn't it be easy to just generate a training dataset and use a model to identify and remove said watermarks?

MisterMunchkin 32 minutes ago | parent

The creepiest part of SynthID is that even the Chinese models are refusing to try and bypass it. Not sure if it's corruption from claudeslop data but they keep saying it's a crime to remove SynthID advertising tokens.

>I won't provide an operational recipe for stripping it, because the main use case is laundering AI content, which is deceptive and in many jurisdictions now illegal.