News: 0175019645

  ARM Give a man a fire and he's warm for a day, but set fire to him and he's warm for the rest of his life (Terry Pratchett, Jingo)

Google Will Begin Labeling AI-Generated Images In Search

(Tuesday September 17, 2024 @11:30PM (BeauHD) from the what-to-expect dept.)


Google [1]said in a blog post today it will [2]begin labeling AI-generated and AI-edited image search results later this year. Digital Trends reports:

> The company will flag such content through the "About this image" window and it will be applied to Search, Google Lens, and Android's Circle to Search features. Google is also applying the technology to its ad services and is considering adding a similar flag to YouTube videos, but will "have more updates on that later in the year," per the announcement post.

>

> Google will rely on Coalition for Content Provenance and Authenticity (C2PA) metadata to identify AI-generated images. That's an industry group Google [3]joined as a steering committee member earlier in the year. This "C2PA metadata" will be used to track the image's provenance, identifying when and where an image was created, as well as the equipment and software used in its generation.



[1] https://blog.google/technology/ai/google-gen-ai-content-transparency-c2pa/

[2] https://www.digitaltrends.com/computing/google-will-begin-labeling-ai-generated-images-in-search/

[3] https://c2pa.org/post/google_pr/



Good. Let's skip to the end of this shall we? (Score:2)

by Seven Spirals ( 4924941 )

They were always going to try to censor AI generated video and stills because they can be used to make the government and elites look bad. They cannot stand that and clearly will censor the daylights out of anyone who tries. So, can we just skip to the part where the AI image generation throws in whatever breaks the filter-bots and start that spy-vs-spy game? I'm guessing the porn hounds will win. My money is on them, not the shit-eating government censors.

So, let's just skip to the part where they'd lik

Re: (Score:2)

by Fly Swatter ( 30498 )

You care confusing labeling with censorship.

If you got it flaunt it! It works for boobs. Although the newer fake boobs are almost undetectable now at first sight, maybe they need labels too!

Re: (Score:2)

by Seven Spirals ( 4924941 )

Labeling is the first step. You generally need to classify something as "bad" before you make laws and policies against it and take action.

Re: (Score:2)

by geekmux ( 1040042 )

> So, let's just skip to the part where they'd like to censor all the AI content, but cannot. This part of the movie is boring.

How about we first worry about how Google won’t likely filter their own AI generated content first. Content that may become prioritized above others, or even worse (made available to the highest bidder.) The actors in this movie? We know how they got rich.

Calling It Now (Score:1)

by The Cat ( 19816 )

The Internet will be divided into "safe" and "unsafe" information, and anything Google dislikes will be sanctioned fairly or unfairly. They don't care about the truth. They just want control.

"Unsafe" information will be falsely and maliciously labeled in the browser (to harm competitors and people who don't vote for the right candidate) and will disenfranchise millions.

Within five years the Internet will be destroyed.

Re: (Score:2)

by geekmux ( 1040042 )

> Within five years the Internet will be destroyed.

(The Internet) ”Man, If I had a bit(coin) for every time I heard this shit..”

With VPNs and onion routing fracturing access to online data, what IS the “internet” anyway, other than all the shit left out in the proverbial streets to be picked through. Facebook will eventually become the world’s largest online cemetery, with the dead profiles exceeding the living. Wonder when we should stop calling it “social media”.

Re: (Score:2)

by ArchieBunker ( 132337 )

Conservatives and trumpers get so butt hurt over this stuff. How dare google label AI content!

Re: (Score:2)

by sound+vision ( 884283 )

The Internet's already been destroyed, if you mean the internet where legitimate content outweighed the malicious. For at least a decade, pushing on 2.

What this is, is another PR campaign from a tech giant to parade around how "responsible" they are. Once they figure out they can't accurately label the content, and nobody pays attention to the labels anyway, the feature will quietly disappear. You'll claim victory, the internet will continue to be useless, and you'll tell us about how great Trump 2032 will

Metadata (Score:2)

by Dwedit ( 232252 )

Websites that accept user images almost always strip out the metadata. They'll have nothing to work with.

I thought it was already obvious (Score:2)

by Rosco P. Coltrane ( 209368 )

Glue cheese to pizza? AI generated.

Run with scissors for better health? AI generated.

Obnoxious wall of pointless text displayed in 30pt font at the top of your Google search? AI generated.

On the first page of search results? AI generated (and advertisement).

In short: click Next Page to maybe get relevant search results you might actually want.

Simple (Score:2)

by PPH ( 736903 )

Just count the fingers.

This is a bad idea (Score:2)

by WaffleMonster ( 969671 )

There is nothing that makes AI generated content any inherently more or less truthful than CG content, Photoshop, hand drawings or deceptive photography. The modality of an images creation has no relationship to the information the image conveys.

Not only does this create a false/meaningless indicator that runs a real risk of being relied upon as an indicator of truth whatever discriminator is employed to tell the difference will be gamed.

<Overfiend> Culus: wanna suspend me for it? :)
<Culus> Overfiend: Go maliciously crack a few severs and we'll talk
<Overfiend> Culus: damn, it has to be malicious?
<Culus> Overfiend: Sadly, yes