There have been a number of good discussions following Chris Messina’s Medium post on “unlisted accounts” and “burners” as new terms of art, including this one at Techdirt. The latter is notable in part for its comment stream, and I thought I’d repost a comment I left there (lightly edited for context) into Medium.
One issue that has become apparent to us in starting Burner — and that is evident in the comment stream on Techdirt — is that we need a vocabulary with more precision around the range of services on the anonymous spectrum. The use of the term “anonymish”, for example, is a pretty good indicator that many services are handwave-y around the concept of anonymity that may or may not be truly, fully anonymous and/or encrypted, and people are either being lazy in their investigations or the services themselves are compromised or being ambiguous.
“Anonymous Coward” comments on Techdirt, for example, are a great example of, basically, an “anonymous guest” mode. It’s very useful and you don’t have to authenticate yourself, but you also don’t get the benefits of an account (e.g., notifications of replies to your comments).
A “Burner” account, both in the sense Chris Messina is talking about in his article and in the sense we think about Burner phone numbers, is an actual account with an actual login, but one under which you can be pseudonymous, and one that you can also easily change if you want to create a new identity (or perhaps maintain multiple identities at once). These services are also great for avoiding finding yourself in marketing databases, or at least “fuzzing” your data within them to some degree. But Burner works by interoperating with the generally available telephone network (CMRS & PSTN carriers). This is its primary advantage, as a single-player user can use it effectively without asking his or her counterparties to download or sign up for anything, but it should be self-evident that any communication through it is only as secure as that entire system —including counterparties’ carriers, hardware, and software — is.
I think of encrypted services as having a different value proposition entirely, but even among them there’s a range (e.g. encrypted message services that still capture metadata, as can be assumed to be the case with companies like Yahoo and Apple who are starting to do encrypted messages, VS companies and services claiming true end-to-end anonymity and encryption). Think of using a service like coinbase vs. buying bitcoins through a strong proxy in a cash transaction. The latter types of services would be the preferred services for whistleblowers, investigative journalists, and fourth-amendment enthusiasts trying to stay truly “off the grid”.
The problem with these kinds of services and the reason I’m not naming any of them—setting aside their potential for nefarious uses—is that you have to get them exactly, perfectly right or risk compromising your system. One social hack or single point of infosec failure could be disastrous, especially so if the vulnerability is invisible to the owners or users of the system, as is often the case in surveillance situations. It’s also easy to misunderstand (or misrepresent) the security level of these systems by some obscure but critically important degree.
As I see it, the rise of “unlisted” publishing and “burner” identities both encourage and symbolize the growing acceptance of non-broadcast (but not private) sharing and the demand for more flexible digital identity options that both warrant deeper consideration.
So please, let’s all consider away. And not to be pedantic, but hopefully it’s obvious that if nuances in the use of these kinds of services are important to you (or your users, readers, etc)—as they definitely should be if you’re using them or encouraging others to—please do your homework.