this post was submitted on 14 Jun 2025
584 points (97.7% liked)

Greentext

6494 readers
862 users here now

This is a place to share greentexts and witness the confounding life of Anon. If you're new to the Greentext community, think of it as a sort of zoo with Anon as the main attraction.

Be warned:

If you find yourself getting angry (or god forbid, agreeing) with something Anon has said, you might be doing it wrong.

founded 2 years ago
MODERATORS
 
top 50 comments
sorted by: hot top controversial new old
[–] rabber@lemmy.ca 8 points 4 hours ago

I like to think humans will go back to interacting with each other in person thanks to AI destroying the internet

[–] nebulaone@lemmy.world 23 points 12 hours ago* (last edited 11 hours ago) (2 children)

Reputation and PGP signatures could be used to verify real human made content. That is, of course, if people actually care, which I think will be rare.

There might be no-ai communities, that require this and are closed down to avoid being scraped for ai training.

Edit: Also AI is already enshittifiying itself, which might get worse if it becomes more widespread than it already is.

[–] alekwithak@lemmy.world 2 points 4 hours ago (1 children)

I like webghost0101's Idea:

[...] a blockchain linked video camera where metadata of footage gets written into the chain to combat fake news and misinformation.

The goal would be to create a proof and record of original footage, to which media publishers and people who share can link towards to verify authenticity/author.

If the media later gets manipulated or reframed you would be able to verify this by comparing it to the original record.

[–] sus@programming.dev 8 points 3 hours ago* (last edited 3 hours ago)

blockchain is a totally useless extra bit glued on there. All the real evidence will be the cryptographic signatures added by the hardware manufacturer (which can be faked, but requires extracting the keys from the "security chip" in the camera which may be very difficult)

all the blockchain does at that point is provide a timestamp of "signed hash of the picture+metadata was uploaded on x date" which can easily be done without blockchain too

[–] prole@lemmy.blahaj.zone 5 points 9 hours ago (3 children)

Why couldn't AI use PGP signatures that suggest they're human?

[–] racketlauncher831@lemmy.ml 1 points 8 minutes ago

Throw the technical bit away. Just think of it as a signature. Yes, an old school, written-with-hand signature.

Does the bank trust me giving you $100 by you having this cheque? Yes. Why? Because I told them what my name is and what my signature is like.

Will the bank give you $100 if you stole my cheque and sign your name on it?

[–] groet@feddit.org 10 points 8 hours ago

Its not about "just having a signature". Its about a web of trust. It only works if you verify if the key belongs to a creator that is actually a person.

Basically creators go to a convention and hand out their public key in person and have other creators sign their key. If you trust creator A is real and they signed the key of creator B, you can have some trust B is also real. And if your buddy went to the convention, met A and B, got their public keys and tells you they are real you can also trust they are real. The more steps/signatures you are away from a creator the less trustworthy they are and nothing really ensures a (human) creator doesn't use AI secretly. If somebody is found to be a fraud everyone has to distrust their key.

[–] PieMePlenty@lemmy.world 5 points 7 hours ago* (last edited 7 hours ago)

Trust is the most important part. You trust someone they made something themselves. They digitally sign their work with a public key that is known to be theirs. You can now verify they (the person you trust) made it.
Once the trusted creator's key is leaked, they are no longer trusted for future works.
AI made content can be freely signed as well, but if you don't trust the origin, the signature doesn't matter anyway since it will just verify it is coming from the AI creator.
The key thing is trust, the signature is just there to verify.

[–] racketlauncher831@lemmy.ml 8 points 10 hours ago (1 children)

Photorealistic porns? What's your problem, man? You have realistic AI and this is all you'll have? Just order a silicon doll and put an AI chip into it! Free ~~sex-sla~~ wife!

[–] Yuki@kutsuya.dev 2 points 1 hour ago

Marketable idea right here

[–] alligalli@feddit.org 7 points 12 hours ago

Time to get up and go outside :)

[–] wowwoweowza@lemmy.world 5 points 12 hours ago (1 children)

Read classics:

Pride and Prejudice" by Jane Austen, "Moby Dick" by Herman Melville, "The Great Gatsby" by F. Scott Fitzgerald, and "1984" by George Orwell.

Start here. There are thousands.

[–] silasmariner@programming.dev 3 points 10 hours ago (2 children)

Why start there with British and US authors? Why not 100 years of solitude, Disgrace, and dream of red mansions?

[–] phoenixz@lemmy.ca 3 points 7 hours ago (1 children)

Because that was what he came up with. It's fine to start with. Your selection is fine too to start with.

[–] silasmariner@programming.dev 1 points 4 hours ago

Not gonna double my response - OP deserved it more - I will say another 'fair enough', give you an upvote, and leave it at that

[–] wowwoweowza@lemmy.world 2 points 7 hours ago (1 children)

For the same reason your user name is not buendiablo?

I guess we can all suffer a little eurocentrism from time to time? But yes — enrich the list with international voices! One of my favorite novels is THE PONDS OF WAGABA by Elichi Amadi… a little known gem any fan of George Eliot would love.

[–] silasmariner@programming.dev 2 points 4 hours ago

Good response, happy with that. Sorry about the implications, I think I just found it a fusty conservative choice. Had it been Infinite Jest and Chaucer I probably wouldn't have bothered responding. Sometimes the idea of 'classics' can seem... narrow and dull. Just wanted to mitigate the notion it was all brown bread (not that I don't love your suggestions tho')

in the futuer we will b fighting the terminators, shotgfun jhon connor

[–] Dogyote@slrpnk.net 8 points 18 hours ago

Simetimes I think the future will resemble the pre-internet era. AI content will be so easy to create that the zone will be flooded with shit, and only a few reputable sources will be trusted, like when there were only a few TV news channels.

[–] StarlightDust@lemmy.blahaj.zone 8 points 18 hours ago

That porn had to be trained on real people's bodies who will never see a penny of it. That's laundered revenge porn.

[–] Rossphorus@lemmy.world 14 points 1 day ago (2 children)

Video evidence is relatively easy to fix, you just need camera ICs to cryptographically sign their outputs. If the image/video is tampered with (or even re-encoded) the signature won't match. As the private key is (hopefully!) stored securely in the hardware IC taking the photo/video, any generated images or videos can't be signed by such a private key.

[–] IlovePizza@lemmy.world 1 points 9 hours ago (1 children)

Wouldn't this be as easy to break as to point a camera at a screen playing whatever you want?

Perhaps not with light field cameras. But then you could probably tamper with the hardware somehow.

[–] sus@programming.dev 3 points 3 hours ago* (last edited 3 hours ago)

getting the picture to perfectly replicate the image on the screen without it being noticeable that it's just a picture of a screen would be so difficult it would probably be easier to modify the camera instead

[–] topherclay@lemmy.world 10 points 23 hours ago (6 children)

So whatever way the camera output is being signed, what's stopping you from signing an altered video with a similar private key and then saying "you can all trust that my video is real because I have the private key for it."

The doubters will have to concede that the video did indeed come from you because it pairs with your key, but why would anyone trust that the key came from the camera step instead of coming from the editing step?

[–] Taleya@aussie.zone 7 points 20 hours ago (2 children)

Mate, digital cinema uses this encryption /decryption method for KDMs.

The keys are tied into multiple physical hardware ids, many of which (such as player/.projector ) are also married cryptographically. Any deviation along a massive chain and you get no content.

Those playback keys are produced from DKDMs that are insanely tightly controlled. The DKDM production itself even more so.

And that's just to play a movie. This is proven tech, decades old. You're not gonna break it with premiere.

[–] tweeks@feddit.nl 3 points 14 hours ago (2 children)

But how would one simple member of the audience easily determine if this whole chain of events is valid, when they don't even get how it works or what to look out for?

You'd have to have a public key of trusted sources that people automatically check with their browser, but all the steps in between need to be trusted too. I can imagine it is too much of a hassle for most.

But then again, that has always been the case for most.

[–] Rossphorus@lemmy.world 1 points 1 hour ago

This is just standard public key cryptography, we already do this for website certificates. Your browser puts a little lock icon next to the URL if it's legit, or provides you with a big, full-page warning if something's wrong with the cert.

[–] Taleya@aussie.zone 3 points 14 hours ago

...what audience?

[–] Valmond@lemmy.world 1 points 12 hours ago (1 children)

This is for restricting use, not proving authenticity of the videos recording. Anyone can spin up keys and sign videos, so in a legal battle it would be worthless.

[–] Taleya@aussie.zone 2 points 12 hours ago* (last edited 8 hours ago) (1 children)

The technology would be extremely easy to adapt, with the certs being tied to the original recording equipment hardware. Given i don't see a $60 ip cam having a dolphin board it would probably be relegated to much higer end equipment, but any modification with a new key would break the chain of veracity

[–] Valmond@lemmy.world 0 points 11 hours ago (2 children)

This is blatantly not true, it would be extremely simple to circumvent. How do you "tie" the cert to a specific hardware without trusting manufacturers? You just can't, it's like putting a padlock on a pizzabox.

[–] Rossphorus@lemmy.world 1 points 1 hour ago

As with everything, trust is required eventually. It's more about reducing the amount of trust required than removing it entirely. It's the same with HTTPS - website certificates only work if you trust the root certificate authorities, for example. Root manufacturer keys may only be certified if they have passed some level of trust with the root authority/authorities. Proving that trust is well-founded is more a physical issue than an algorithmic one. As it is with root CAs it may involve physical cybersecurity audits, etc.

[–] Taleya@aussie.zone 2 points 8 hours ago (1 children)

I literally explained earlier how this exact technology is used in digital cinema dude c'mon.

[–] Valmond@lemmy.world 1 points 7 hours ago* (last edited 7 hours ago)

That doesn't mean it's useful for forensics, IMO.

Edit: not saying it wont be though, just that it's not as bullet proof as you'd think, IMO.

load more comments (4 replies)
load more comments
view more: next ›