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

YouTube now requires to label their realistic-looking videos made using AI

blog.google

Post reply on HN
Print view View original

421–430 of 517 posts

Re: YouTube now requires to label their realistic-looking videos made using AI

#421

Post by Cthulhu_ »

JohnFen wrote: post #239
jjcm wrote: post #45

Earlier quoted context omitted.

One black hat thing I'm curious about though is whether or not this tag can be weaponized. If I upload a real event and tag it as AI, will it reduce user trust that the real event ever happened?

I fear that we're barrelling fast toward a future when nobody can trust anything at all anymore, label or not.

And this isn't new. A fad in films in the 90's was hyper-realistic masks on the one side, and make-up and prosthetics artists on the other, making people look like other people.

Faking things is not new, and you've always been right to mistrust what you see on the internet. "AI" technology has made it easy, convenient, accessible and affordable to more people though, beforehand you needed image/video editing skills and software, a good voice mod, be a good (voice) actor, etc.

Re: YouTube now requires to label their realistic-looking videos made using AI

#422

Post by Cthulhu_ »

sverhagen wrote: post #381
JohnFen wrote: post #239

Earlier quoted context omitted.

I fear that we're barrelling fast toward a future when nobody can trust anything at all anymore, label or not.

I was immediately thinking that the #AI labels are going to give people a false sense of trust, so that when someone posts a good-enough fake without the #AI label, it can do damage if it goes viral before it gets taken down for the mislabeling. (Kudos for the effort, though, YouTube.)

Behind the scenes, I'm 99% confident that Google has deployed AI detection tools and will monitor for it.

That said, unless all the AI generators agree on a way to add an unalterable marker that something is generated, at one point it may become undetectable. May.

Re: YouTube now requires to label their realistic-looking videos made using AI

#423

Post by sumedh »

Devasta wrote: post #153

I hope this allows me to filter them entirely. If it wasn't worth your time creating it, its not worth my time looking at it. I am generally very skeptical of these tags though, I suspect a lot of them are in place to stop an AI consuming its own output rather than any concern for the end user.

> If it wasn't worth your time creating it

Once something like Sora is available to the public, its going to be game over. A new bunch of creators will use it to "create" videos and I am sure you will change your mind then.

Re: YouTube now requires to label their realistic-looking videos made using AI

#424

Post by pjc50 »

alphazard wrote: post #188
the_duke wrote: post #31

They don't bother to mention it, but this is actually to comply with the the new EU AI act. > Providers will also have to ensure that AI-generated content is identifiable. Besides, AI-generated text published with the purpose to inform the public on matters of public interest must be labelled as artificially generated. This also applies to audio and video content constituting deep fakes https://digital-strategy.ec.eu…

Is anyone else worried about how naive this policy is? The solution here is for important institutions to get onboard with the public key infrastructure, and start signing anything they want to certify as authentic. The culture needs to shift from assuming video and pictures are real, to assuming they are made the easiest way possible. A signature means the signer wants you to know the content is theirs, nothing else…

PKI has been around for, what, 30 years? Image authentication is just not going to happen at this point, because everyone's got too used to post-processing and it's a massive hassle for something that ultimately doesn't matter because real people use other processes to determine whether things are true or not.

Re: YouTube now requires to label their realistic-looking videos made using AI

#425

Post by astro- »

I’m wondering whether another motivation for this could be trying to keep the data set as clean as possible for future model training.

Creating videos takes quite a bit of time. If AI video generation becomes widely available, pretty soon, there could be more AI content being uploaded to YouTube than human-made stuff.

Presumably, training on AI generated stuff magnifies any artefacts/hallucinations present in the training set, reducing the quality of the model.

Re: YouTube now requires to label their realistic-looking videos made using AI

#426

Post by LtWorf »

eek2121 wrote: post #297
miki123211 wrote: post #282

Earlier quoted context omitted.

also the three letter agencies (not just from the US) will have access to private keys of at least some manufacturers, allowing them to authenticate fake events and sow chaos by strategically leaking keys for cameras that recorded something they really don't like.

For all the folks that bash the United States for "reasons" this one gave me a chuckle. Our handling of privacy and data and such is absolute ass, but at least we *can* hide our data from big government with little repercussion in most cases (translation: you aren't actively being investigated for a crime that a judge isn't aware of) Of course that says nothing about the issues of corruption of judges in the court sy…

> but at least we can hide our data from big government with little repercussion

They come and ask. You say no? They find cocaine in your home.

You aren't in jail because you refused to hand out data. You are in jail because you were dealing drugs.

Re: YouTube now requires to label their realistic-looking videos made using AI

#427

Post by A1kmm »

8note wrote: post #360
dataflow wrote: post #348

Earlier quoted context omitted.

What does this have to do with root? The camera chip would be the one signing the data flowing through it, not the Android kernel.

If you do a jpeg compression, or crop the file, then does that signature matter anymore?

Cryptography also has answers for some of this sort of thing. For example, you could use STARKs (Succinct Transparent Arguments of Knowledge) to create a proof that there exists a raw image I, and a signature S_I of I corresponding to the public key K (public input), and that H_O (public input) is a hash of an image O, and that O is the output of providing a specified transformation (cropping, JPEG compression) to I.

Then you give me O, I already know K (you tell me which manufacturer key to use, and I decide if I trust it), and the STARK proof. I validate the proof (including the public inputs K and H_O, which I recalculate from O myself), and if it validates I know that you have access to a signed image I that O is derived from in a well-defined way. You never have to disclose I to me. And with the advent of zkVMs, it isn't even necessarily that hard to do as long as you can tolerate the overhead of running the compression / cropping algorithm on a zkVM instead of real hardware, and don't mind the proof size (which is probably in the tens of megabytes at least).

Re: YouTube now requires to label their realistic-looking videos made using AI

#428

Post by krapp »

int_19h wrote: post #379
krapp wrote: post #298

Earlier quoted context omitted.

Calling it an error implies the model should be expected to be correct, the way a calculator should be expected to be correct. It generates syntactically correct language, and that's all it does. There is no "calculation" involved, so the concept of an "error" is meaningless - the sentences it creates either only happen to correlate to truth, or not, but it's coincidence either way.

That's one hell of a coincidence if it just "happens" to write syntactically correct code that does what the user asked, for example.

It is.

It's a language model, trained on syntactically correct code, with a data set which presumably contains more correct examples of code than not, so it isn't surprising that it can generate syntactically correct code, or even code which correlates to valid solutions.

But if it actually had insight and knowledge about the code it generated, it would never generate random, useless (but syntactically correct) code, nor would it copy code verbatim, including comments and license text.

It's a hell of a trick, but a trick is what it is. The fact that you can adjust the randomness in a query should give it away. It's de rigueur around here to equate everything a human does with everything an LLM does, including mistakes, but human programmers don't make mistakes the way LLMs do, and human programmers don't come with temperature sliders.

Re: YouTube now requires to label their realistic-looking videos made using AI

#430

Post by cseleborg »

This is interesting because it highlights the trust we've always placed in real-looking images. It brings real-looking images down to the same level as text.

It's always been possible to write fake news. We've never had to add disclaimers at the top of textual content, e.g. "This text is made to sound like it describes real events, but contains invented and/or inaccurate facts." We feel the need to add this to video because until now, if it looked real, it probably was (of course, "creative" editing has existed for a long time, but that's still comparatively easy to spot).

It's the end of a media era, really.

Post reply on HN