Conversation

Replying to

@matthew_d_green

Apropos of nothing, what would be gained?

This Tweet was deleted by the Tweet author. Learn more

Replying to

and

It’s almost as though Google wants stolen emails to be cryptographically authenticated.

Replying to

and

So are we saying that we would like the authenticity of OLD email to be completely disavowed? This would then seem to provide the same protection to historical senders as per OTR Messaging, is that an advantage?

Replying to

and

DKIM is supposed to be an anti-spam mechanism. Not a forever-non-repudiable-cryptographic-authentication mechanism that encourages and facilitates email leaks.

Replying to

and

There’s no reason for anyone to be able to say “yes this three year old message was definitely signed by Google”.

Replying to

and

But you can't prevent that. Someone will log old keys.

Replying to

and

Logging old public keys is inevitable. But when new collections of archived stolen emails are leaked on the Internet, they become deniable if the secret keys are published.

Replying to

and

Ah, excellent! I never thought of intentionally burning your secret keys as a non-repudiation mitigation! I love it.

Replying to

Old keys need to be published preemptively though. If it's only after a leak, leaker could use a notary service long before they publicly leak.

Replying to

Could regularly rotate DKIM keys and then publish old keys a couple weeks after they're removed from the DNS records. It still leaves a period of time where DKIM can be used as proof that the email was sent though. Within that window, they can use something like OpenTimestamps.

11:03 PM · Nov 1, 2020Twitter Web App

Replying to

It’s definitely not ideal. But it’s the difference between capturing someone’s email in real time and getting access to their archived mailbox. In the real world the latter scenario seems much more common.