The European Comission is looking for feedback on forcing retention of metadata from all communication services for “a reasonable period of time”, for purposes of criminal investigation!
Which means encrypted messaging without a backdoor would be illegal if this passes! That’s a slippery slope!
That basically means an attacker with some skill could read any data from anyone (correct me if I’m wrong but I think you can infer the content from the metadata in 90% of cases)
For more detail on why it’s bad, click the link below and read literally any feedback comment.
Go ahead and give some feedback! You can do so even if you are not an EU citizen!
https://ec.europa.eu/info/law/better-regulation/have-your-say/initiatives/14680-Impact-assessment-on-retention-of-data-by-service-providers-for-criminal-proceedings-/_en
@soatok @echo_pbreyer @privacy @technology
#Europe #privacy #encryption
Metadata is not content, so no E2E would not be illegal. Metadata is things like who sent messages to who at what time, duration, volume of data, other externally parsable metrics like that.
True. Also, it would be helpful to actually understand what kind of metadata is this referring to and specifically in which cases does this apply and which cases are exempt… because I expect that if the design of a service explicitly makes it so all the metadata you can collect is not helpful/reliable, then you wouldn’t be forced to redesign the service, you’ll just provide metadata that’s unreliable.
I feel these kind of measures never are really effective at stopping organized criminal activity (since those looking for a way will find it), what they are effective at is tracing/tracking non-criminal private use.
I would say no criminal uses public services to do their business, but then there was the whole Signal thing at the DOD…
@ShellMonkey That’s true, but these should also be encrypted, don’t you think?
Sure the messages could still be encrypted but from the metadata you can most of the time infer the content.
It’s a bit long, but you can read this if you’re interested
https://www.pnas.org/doi/full/10.1073/pnas.1508081113
It depends on how many layers of obfuscation you are looking to deal with. There always needs to be some publicly shared token to initiate a connection, even if that’s only the public key of a asymmetric key pair to a 3rd party auth system.
There are ways to do it, but part of the difficulty is there are so many ways to do it that coming to an agreed method is like herding cats.
@ShellMonkey Be careful, obfuscation isn’t encryption!
And no, there doesn’t need to be a publicly shared token! Take a look at how simplexchat does it!
https://simplex.chat/
Right, I just use the term sometimes to say hiding things, even if it’s hidden via encrypting it.
Will have to delve into the papers for simplex later here, but in the end there needs to be some type of known identity to pin a communication to, otherwise you’ve already breached the confidentiality point of the security triad by not authenticating the recipient.