Matthew Green, associate professor of computer science at Johns Hopkins University in the US, wants Google and other email providers to make it possible for people to deny they’ve written old email messages.
He has asked the Gmail goliath, as the largest commercial email service, to rotate its Domain Keys Identified Mail (DKIM) encryption keys periodically and to publish old keys to reduce the incentive for hackers to steal and leak email messages.
In so doing, he’s advocating for a “right to be forgotten” for aged email, or more precisely for the ability to plausibly deny authorship of messages published without permission, after a certain period of time.
In an online post published earlier this week, Green argues that DKIM has become “a monster.”
As he explains, email service providers like Google’s Gmail add a DKIM cryptographic signature to email messages. Used in conjunction with other email security protocols like DMARC, which can prevent cryptographic signatures from being removed, message recipients can be assured of the authenticity of signed messages and their contents.
But Green is concerned that the ability to verify message authenticity incentivizes hackers to steal email. DKIM, he argues, was designed to fight spam, not to tie people to their messages forever.
“This new non-repudiation feature was not part of DKIM’s design goals,” Green says. “The designers didn’t intend it, nobody discussed whether it was a good idea, and it seems to have largely taken them by surprise. Worse, this surprise feature has some serious implications: it makes us all more vulnerable to extortion and blackmail.”
It also enables reporters to authenticate leaked email messages, like those published by Wikileaks in 2016 from former Obama administration counselor John Podesta, by news organizations in 2017 covering messages sent by President Trump’s personal lawyer Mark Kasowitz, and by others reporting on matters of public interest. As Green points out, the Associate Press offers a shell script to assist investigative reporters interested in verifying DKIM signatures.
Nonetheless, Green contends DKIM’s unintended side effect of permanent accountability should be rolled back. He says that while people may find the consequences agreeable “because it suits a partisan preference, or because the people who got ‘caught’ sort of deserved it,” everyone is potentially vulnerable to being victimized.
“[B]ad things happen to good people too,” he says. “If you build a mechanism that incentivizes crime, sooner or later you will get crimed on.”
If Google were to publish its DKIM keys after a certain period of time, then messages signed with those decommissioned keys could no longer be convincingly tied to a given author.
That’s not obviously a desirable outcome at a time when lack of accountability for online communication has made misinformation and disinformation a major concern around the world.
Green allows that cryptographic authentication can be useful in some circumstances but points out that no commercial email customer has asked for DKIM as a default feature. If people want to author messages that can be cryptographically linked to them, they can choose to use tools like GnuPGP.
As an alternative to DKIM, Green and two colleagues, Michael A. Specter and Sunoo Park, have proposed a way to create cryptographic signatures that expire, becoming forgeable after a certain period of time without invalidating the associated public key for unexpired signatures.
The Register asked Google whether it would consider rotating and publishing its DKIM keys. We’ve not heard back. ®