Like the title states looking for E2EE apps (Android and iOS) without going into much details or needs to be robust enough and easy to use for anyone and stable for operations that are susceptible to constant electronic warfare. I did some research and thought about replacing Signal with Molly and wondering if it will still work if Signal leaves the EU, but am also worried about its updates to patch vulnerabilities in a timely manner. I appreciate the help I am a “Jack of all trades and master of none” when it comes to these types of programs, but am also the go to currently in my unit since I am somewhat knowledgeable about exploits and attacks that can compromise systems would be great if there was an desktop as well (like Signal) and would also be nice if it was FOSS and auditable ( I know that’s kind of redundant ) I know it’s a tall order to ask but figured I would try. I really appreciate the help so much and hope I did things by the rules here and don’t get flamed if this has already been covered ( I searched but my skills with searching the fediverse is low
It’s not encryption, all goes end-to-end. They will force app clients to add a “leak” request that sends the hash of each image you send to compare if it matches with child porn. It’s explained on another post on Lemmy and it looks so hard and so impossible to be implemented that I doubt it will actually work. The chat is still end-to-end.
They want to also check them with ai. Hash alone would be bad. But ai is worse. Ya got/are young looking gf. Well if ya send nudes some cop will most likely see your nudes if chat controll really comes.
Source: the new law proposal
These cops probably
https://www.euronews.com/2023/08/07/nazi-symbols-and-child-pornography-found-in-german-police-chats
Yes.
They will check their own images and police themselves lol (actually there will be an extra committee for this so just joking)
AI is the solution to all our problems… /s
Do you have the source link of this pls? thx!
https://eur-lex.europa.eu/resource.html?uri=cellar:13e33abf-d209-11ec-a95f-01aa75ed71a1.0001.02/DOC_1&format=PDF
Here. On page 52, Article 10 3.a
They are explicitly talking about known or new material. Even though they don’t state the technology, AI is the only possible one (maybe there are more but they WILL have the same issue, ai has)
They also go indepth in a centralized db, where all this shit will be stored, to retrain this model.
Yea it is fucked up.
But AI is too expensive to use it on any text sent by any European citizen/bot.
I would guess it rather refers to images. But it doesn’t matter if it is too expensive. Ai is the only thing that can do the stuff they want.
Are you sure? Using AI for any text and image that any European citizen (742,083,786 people if I’m not wrong) is pretty heavy… They will need to spend too much money for AI usage if this is the stuff they want or the only thing they can do.
EDIT: Sorry, if you mean only about images is still very heavy, no AI needed here, they said it’s a client side implementation and still sending random hashes… any hacker can just send random hashes and block database request with a DoS… and they will get many false positives… Not viable.
O God, the hash one isn’t bad, just stop there for the sake of all that is holy
The hash one has the one issue of you could simply put political shit in the db and find out your political opponents, but the hash one is debatebal.