Meta Will Start Labeling AI-Generated Images On Facebook, Instagram and Threads (reuters.com) 15
Meta will start detecting and labeling images generated by artificial intelligence from Google and OpenAI's services, among others. According to Reuters, the company will use a set of invisible markers built into the files to help with the identification process. From the report: Meta will apply the labels to any content carrying the markers that is posted to its Facebook, Instagram and Threads services, in an effort to signal to users that the images -- which in many cases resemble real photos -- are actually digital creations, the company's president of global affairs, Nick Clegg, wrote in a blog post. The company already labels any content generated using its own AI tools. Once the new system is up and running, Meta will do the same for images created on services run by OpenAI, Microsoft, Adobe , Midjourney, Shutterstock and Alphabet's Google, Clegg said.
Suppression! (Score:1)
Help! They're suppressing my free speech by letting everyone see what I posted but telling them it's not real. Everyone on the planet can still see what I posted, but they're suppressing me! They can't do that! Who are they to determine what's real on their site for which I pay nothing to use? They're colluding with all those other companies to suppress what I can do! I was entrapped!
Did I get them all?
Re: (Score:2)
Re: (Score:2)
If they're working with the AI companies it should be (relatively) easy to verify if they generated one of the pictures. Or, the person themselves can show proof they didn't use AI.
Not saying it won't happen, but if that's the worst that happens, it's insignificant in the grand scheme of things since nothing is being removed.
Re: (Score:2)
If they're working with the AI companies it should be (relatively) easy to verify if they generated one of the pictures.
That would be funny.
I can see the AI companies charging for that stuff.
Fail (Score:5, Insightful)
By identifying the marked images, they're going to be implying the unidentified ones aren't AI-generated, when they're merely not AI-marked.
Anyone who wants to deceive will actually get a boost to their efforts.
Helps out the bad guys (Score:1)
The bad guys will be the ones who bother to remove the watermarks
Re: (Score:2)
It's easy to make watermarks that can't be removed without ruining the picture.
It's also easy, and probably universally done by "the bad guys" to run your own image generator that doesn't include watermarks.
Re: (Score:2)
They will not be the only ones to get a boost. Now real images can be censored by labeling them as AI-generated.
Ah, so you have to pay extra... (Score:2)
Re: (Score:2)
Bypassing? (Score:1)
Will Photoshop mark their AI too? (Score:2)
Of course it won't, despite all their "generative AI" and "content-aware fill" features. Having only some sources of AI content marked just gives a false sense of security. It's not going to take long for the invisible marking system to be reverse engineered and removal techniques and software will quickly become available. Whack-a-mole here we come.
In the US, presumably the first amendment is going to make it tricky to try to force AI vendors to apply invisible marking, especially with open-source softw
The evil flag (Score:2)
The Meta needs a filter (Score:1)