Google Search will soon label images as AI-generated, edited with photo editing software or if it was taken with a camera in the image search results. This label will be added to the about this image feature, according to The Verge who spoke to Laurie Richardson, vice president of trust and safety at Google.
No, you cannot use metadata as even extremely weak evidence that an image is real. It is less than trivial to fake, and the second anyone even hints at making it a standard approach, it will be on every photo anyone uses to mislead anyone.
Most photos on the internet are camera phones, and you absolutely are not entitled to know what phone someone has. Knowing someone’s phone has infinitely more value to fingerprinting a user than including metadata could ever theoretically have to demonstrate whether a photo is legitimate or not.
Photos without a specific, on record provenance from a credible source are no longer useful for evidence of anything. You cannot go back from that.
Meta data creates a string, if you want to claim ownership of an image and I show an image with earlier metadata, who’s is the real one? Yes it can be faked, but it can also be traced. Thats not a reason to not do something, the hell? That’s like suggesting you can’t police murders because someone can fake a murder.
What is identifiable about the type of phone you have…? Anyone that sees you in public has that information lmfao, there’s far more “fingerprintable” data in the exif than the device that anyone can visually see you have…… that’s the strangest privacy angle I’ve seen and you’re talking like it’s this big huge issue? I’ve asked you to explain and you haven’t, why is this?
And without that exif data you can’t prove any of that… you realize this… yeah…?
What is your point here? That you’re concerned that you might have someone knowing your phone? You realize you can scrub that information yourself if you’re not worried about proving authenticity…? Yeah…?
You very clearly have no idea whatsoever what you’re talking about. This is all complete nonsense.
Anyone can write exif data to say anything they want it to. You “showing an image with earlier metadata” is completely arbitrary and doesn’t tell anyone literally anything about which one is more likely to be “real”. Again, it’s not “weak” or “bad” evidence. It is literally not capable of being evidence.
So you gonna address what’s identifiable about a phone… or are you just gonna ignore this and scream about the one thing we know can prove authenticity of an image? I’ve addressed the can be faked… you gonna address any of my points…?
I said I had a little knowledge, do you have a point here or you just gonna scream that exif data can be faked? I was trying to have a civil conversation about this.
If there’s an image with two different exifs data, this will flag it, problem solved, what’s your issue…? Isn’t that the point? Flag fake images…?
What device you use is one of the biggest data points advertisers and trackers use to fingerprint you across the internet. No, “I use a Google Pixel 9” does not, by itself, de-anonymize you, but it does make a big dent when combined with other information.
You keep talking about “proving the authenticity of an image” with something that does not even move you .00000001% towards an image being legitimate. It is literally zero information about that question in every possible context. It is, eventually, if you throw out every camera on the planet and use heavy cryptography, theoretically possible to eventually, in the future, provide some evidence that some future picture came from some specific camera, but it will still not be proof that what that camera processed wasn’t manipulated.
A group of tech companies created the C2PA system beginning in 2019 in an attempt to combat misleading, realistic synthetic media online. As AI-generated content becomes more prevalent and realistic, experts have worried that it may be difficult for users to determine the authenticity of images they encounter. The C2PA standard creates a digital trail for content, backed by an online signing authority, that includes metadata information about where images originate and how they’ve been modifie
For 5 fucking years already….
Okay, what does an image metadata and advertising have to do with each other…? I’m not here for conspiracy theories, I’m here to have a discussion, which you clearly can’t do.
You claim I don’t know much… I stated as much… yet you don’t know how images are verified …? The fuck…? Go off on whatever tangent you want, but exit data is the only way to determine if a photo is legitimate… yes it can be faked… congrats for pointing that out and only that this entire time… even though I already mentioned that…
What’s your point dude? Seriously I’m blocking you if you can’t have a discussion. Proof of ownership and detecting fakes are two mutually inclusive things, they can both be used to help the others legitimacy, why are you only looking at this from one angle here? Exif is for ownership, the methods in the comment I responded to are for other things. I mentioned THIS previously as well….
No, you cannot use metadata as even extremely weak evidence that an image is real. It is less than trivial to fake, and the second anyone even hints at making it a standard approach, it will be on every photo anyone uses to mislead anyone.
Most photos on the internet are camera phones, and you absolutely are not entitled to know what phone someone has. Knowing someone’s phone has infinitely more value to fingerprinting a user than including metadata could ever theoretically have to demonstrate whether a photo is legitimate or not.
Photos without a specific, on record provenance from a credible source are no longer useful for evidence of anything. You cannot go back from that.
Meta data creates a string, if you want to claim ownership of an image and I show an image with earlier metadata, who’s is the real one? Yes it can be faked, but it can also be traced. Thats not a reason to not do something, the hell? That’s like suggesting you can’t police murders because someone can fake a murder.
What is identifiable about the type of phone you have…? Anyone that sees you in public has that information lmfao, there’s far more “fingerprintable” data in the exif than the device that anyone can visually see you have…… that’s the strangest privacy angle I’ve seen and you’re talking like it’s this big huge issue? I’ve asked you to explain and you haven’t, why is this?
And without that exif data you can’t prove any of that… you realize this… yeah…?
What is your point here? That you’re concerned that you might have someone knowing your phone? You realize you can scrub that information yourself if you’re not worried about proving authenticity…? Yeah…?
You very clearly have no idea whatsoever what you’re talking about. This is all complete nonsense.
Anyone can write exif data to say anything they want it to. You “showing an image with earlier metadata” is completely arbitrary and doesn’t tell anyone literally anything about which one is more likely to be “real”. Again, it’s not “weak” or “bad” evidence. It is literally not capable of being evidence.
So you gonna address what’s identifiable about a phone… or are you just gonna ignore this and scream about the one thing we know can prove authenticity of an image? I’ve addressed the can be faked… you gonna address any of my points…?
I said I had a little knowledge, do you have a point here or you just gonna scream that exif data can be faked? I was trying to have a civil conversation about this.
If there’s an image with two different exifs data, this will flag it, problem solved, what’s your issue…? Isn’t that the point? Flag fake images…?
What device you use is one of the biggest data points advertisers and trackers use to fingerprint you across the internet. No, “I use a Google Pixel 9” does not, by itself, de-anonymize you, but it does make a big dent when combined with other information.
You keep talking about “proving the authenticity of an image” with something that does not even move you .00000001% towards an image being legitimate. It is literally zero information about that question in every possible context. It is, eventually, if you throw out every camera on the planet and use heavy cryptography, theoretically possible to eventually, in the future, provide some evidence that some future picture came from some specific camera, but it will still not be proof that what that camera processed wasn’t manipulated.
….
https://arstechnica.com/information-technology/2024/09/google-seeks-authenticity-in-the-age-of-ai-with-new-content-labeling-system/
Its literally the method that’s used…
For 5 fucking years already….
Okay, what does an image metadata and advertising have to do with each other…? I’m not here for conspiracy theories, I’m here to have a discussion, which you clearly can’t do.
You claim I don’t know much… I stated as much… yet you don’t know how images are verified …? The fuck…? Go off on whatever tangent you want, but exit data is the only way to determine if a photo is legitimate… yes it can be faked… congrats for pointing that out and only that this entire time… even though I already mentioned that…
What’s your point dude? Seriously I’m blocking you if you can’t have a discussion. Proof of ownership and detecting fakes are two mutually inclusive things, they can both be used to help the others legitimacy, why are you only looking at this from one angle here? Exif is for ownership, the methods in the comment I responded to are for other things. I mentioned THIS previously as well….