Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

SynthID: A tool to watermark and identify content generated through AI

deepmind.google

Post reply on HN
Print view View original

51–60 of 81 posts

Re: SynthID: A tool to watermark and identify content generated through AI

#51

Post by mikestorrent »

littlestymaar wrote: post #48
nightski wrote: post #42

Earlier quoted context omitted.

How would you prove that something was generated by AI yet did not include a watermark?

You can trivially enforce that at the AI provider level, which covers 99% of the problem the law is designed to address. Of course it doesn't cover the issue of foreign state psyop operations but the fact that enforcing laws against organized crime and adversary state actors is hard isn't specific to AI.

Are you not aware of open-weights models and local generation? I think the vast majority of deepfake content is being genned in basements on RTX cards, not on public providers. People already have all this content, and have archives of it, and can run it airgapped. Cat is out of bag.

Re: SynthID: A tool to watermark and identify content generated through AI

#52

Post by manbash »

It's nice that they explain the "what" (...it is doing) but not the "why". Who is going to use it and for what reasons?

Also, if it's essentially a sort of metadata, can't the output generated image be replicated (e.g. screenshot) and thus stripped of any such data?

Re: SynthID: A tool to watermark and identify content generated through AI

#53

Post by u1hcw9nx »

raincole wrote: post #32
parliament32 wrote: post #30

Note that watermarking (yes, including text) is a requirement[1] of the EU AI Act, and goes into effect in August 2026, so I suspect we'll see a lot more work in this space in the near future. [1] Specifically, "...synthetic audio, image, video or text content, shall ensure that the outputs of the AI system are marked in a machine-readable format and detectable as artificially generated or manipulated", see https://a…

EU really like unenforceable regulations, doesn't it?

It's regulation for providers, so it's easy to enforce.

Re: SynthID: A tool to watermark and identify content generated through AI

#56

Post by osculum »

gregorkas wrote: post #6

I genuinely feel that in this AI world we need the inverse. That every analogue or digital photo taken by traditional means of photography will need to be signed by a certificate, so anyone can verify its authenticity.

Years ago, I worked at Apple at the same time as Ian Goodfellow. This was before ChatGPT (I'd say around 2019).

I had the chance to chat with him, and what I remember most was his concern that GANs would eventually be able to generate images indistinguishable from reality, and that this would create a misinformation problem. He argued for exactly what you’re mentioning: chips that embed cryptographic proof that a photo was captured by a camera and haven't been modified.

Re: SynthID: A tool to watermark and identify content generated through AI

#58

Post by themafia »

mikestorrent wrote: post #49
pegasus wrote: post #45

Earlier quoted context omitted.

Yeah, they also outlawed murder. And stealing. And bribing officials. All universally unenforceable. Weird...

Well, consider the case with murder: they're not demanding that people proactively implement a system to prevent it from happening, are they? You're just not allowed to do it, in the sense that the system will attempt to find you, prove your guilt, and punish you after the fact.

> they're not demanding that people proactively implement a system to prevent it from happening, are they?

What do you think a "background check" is?

Re: SynthID: A tool to watermark and identify content generated through AI

#59

Post by bicx »

themafia wrote: post #58
mikestorrent wrote: post #49

Earlier quoted context omitted.

Well, consider the case with murder: they're not demanding that people proactively implement a system to prevent it from happening, are they? You're just not allowed to do it, in the sense that the system will attempt to find you, prove your guilt, and punish you after the fact.

> they're not demanding that people proactively implement a system to prevent it from happening, are they? What do you think a "background check" is?

Definitely not murder prevention

Re: SynthID: A tool to watermark and identify content generated through AI

#60

Post by pegasus »

mikestorrent wrote: post #49
pegasus wrote: post #45

Earlier quoted context omitted.

Yeah, they also outlawed murder. And stealing. And bribing officials. All universally unenforceable. Weird...

Well, consider the case with murder: they're not demanding that people proactively implement a system to prevent it from happening, are they? You're just not allowed to do it, in the sense that the system will attempt to find you, prove your guilt, and punish you after the fact.

I imagine it would be the same for making (use of) models which don't add these watermarks, no? The punishable crime is providing or using the service.
Post reply on HN