January 21, 2016

Action Item: Protecting Ourselves from Encryption Backdoors

Yesterday, in "The Politicians' Encryption Backdoor Fantasies Continue -- and Legislating Pi" ( https://lauren.vortex.com/archive/001147.html ), I discussed moves in the U.S. Senate to convene a commission to proceed toward their fantasy goal of finding a way to backdoor strong encryption algorithms "while still protecting the privacy of honest users."

As I noted then, this is an impossible task, since the very act of building backdoors into these algorithms (ostensibly for law enforcement and intelligence needs) would make these encryption systems exceedingly vulnerable both to "official" abuse and vast third-party black-hat hacking attacks -- including by terrorist groups and other criminals -- who of course for themselves will continue using easily available strong crypto systems without backdoors.

I viewed the call for an encryption commission to be essentially a smokescreen for moving toward the government's ultimate goal -- being able to read all encrypted communications upon demand.

Within hours of my posting yesterday came word that there's already a bipartisan move in the senate to not bother with any commission, to just move directly to legislation mandating law enforcement access to encrypted communications. Period.

I rest my case -- smokescreen proven. Q.E.D.

Whether or not such legislation passes immediately is not really the point, because ultimately the odds are very high that sooner or later something like it will become law here in the U.S. -- and likely in many other countries as well. Not just the obvious suspects like Russia and China, but in the EU also, which constantly speaks out of both sides of its mouth when it comes to privacy and surveillance issues.

So sometime soon -- be it one year, or two, perhaps a bit more if such laws become entangled in court cases (as seems likely), we will be facing the reality of strong, end-to-end encryption essentially being outlawed, at least in the context of the major Internet services that most of us depend upon.

These are the firms that government is currently most concerned about -- Google, Apple, Microsoft, and more -- who have been moving rapidly and correctly to provide their users with strong crypto (e.g. on smartphones) that even the firms themselves can't crack. Such moves have been triggered in large part by the continuing parade of government overreach when it comes to accessing the data in these devices.

Also, these same services have been moving toward providing stronger crypto for their centralized "cloud" services as well, including "only the user holds the keys" encrypted file/data storage systems.

All of these services and more will likely be targeted by government encryption backdoor legislation in coming months and years.

The question is, what are we going to do about it?

Or first, a different question.

Do we care?

The pro-backdoor argument runs something like this ...

Bad guys use encryption (to some extent not clearly known, but expanding). Government can't monitor their communications to prevent or solve terrorist attacks or other crimes (child pornography is frequently mentioned in the latter category) without access to that data. The risks and potential loss of privacy that honest users face from backdoors in these systems for law enforcement and intelligence use is the price we have to pay for living in a 21st century society.

If you're in the category just described, you likely need not read any further in this post.

The counter-argument is that serious bad guys will quickly move (if they haven't already) to crypto systems that don't have backdoors, leaving mainly honest users on the compromised systems.

Encryption experts and computer scientists are in virtually unanimous agreement that any attempts to backdoor these systems weakens them in fundamental ways, making them massively vulnerable not only to government abuse and demonstrated ineptitude (such as permitting the personal info of millions of persons to be obtained by crooks from government computers), but also hacking attacks of all sorts, including by criminal gangs and worse. With so much of our financial and personal information now online -- whether we all like that or not -- purposely weakening encryption systems for honest users is intolerable.

If you're in this second camp -- as am I -- we're back to the "What do we do about it?" question.

And actually, the answer is quite clear. Data that is already encrypted when it is stored or shared, using strong encryption systems that are validated to not contain backdoors (a much tougher validation task than laymen might assume) is not subject to the sorts of backdoor snooping or backdoor hacking exploitations as would be data encrypted on systems mandated to contain backdoors.

Perhaps even more to the point, government still has ways to target particular criminals or other evildoers when they really need to -- in particular through "endpoint" surveillance of various sorts directly on targeted PCs. But generally speaking, backdoored centralized crypto systems represent much greater risks related to mass abuse, mass hacking, and mass surveillance. And this holds true irrespective of how "clever" proponents try to be about splitting up encryption keys and the various related key handling processes.

So honest, good users who feel that they deserve at least the same level of encryption protection as bad, evil users will need to be ramping up their own use of strong encryption systems locally, so data that doesn't need to be stored unencrypted in central services for processing is encrypted in ways that backdoors cannot typically penetrate.

Which data will fall into this category will be largely an individual choice, of course. Cloud environments provide immense value in a vast number of ways -- email systems, file searching, document creation and editing -- on and on. Most of these -- given current tech, anyway -- require that data be unencrypted in the cloud so that it can be processed for the user. On the other hand, for end-to-end communications -- say from one phone user to another, or between users in various other contexts, the need for central processing of those messages -- other than passing them along encrypted as they are -- will often be nil. So central systems in these circumstances become the conduits of data that they do not need to decrypt nor interpret.

A bitter irony is that while some terrorist groups seem to have all manner of sophisticated and relatively standardized strong encryption systems that government backdoors are unlikely to reach, ordinary honest users are faced with a confusing hodgepodge of crypto systems that are generally hard to use, often incompatible, and basically just a pain in the neck that discourage their widespread adoption, especially by non-techies.

The relatively straightforward bottom line?

Given the quite reasonable assumption that mandated encryption backdoors legislation targeting large Internet services is very likely coming -- exact timing unclear, but on the way -- efforts need to be expanded right now toward making personal encryption systems that can run on users' local computers as simple, reliable, automatic, and ubiquitous as possible.

Not to shield evil. Not to mask criminals and terrorists.

But simply to protect the good guys. The rest of us. You and me.

And that's a fact.

Be seeing you.

I have consulted to Google, but I am not currently doing so -- my opinions expressed here are mine alone.

Posted by Lauren at January 21, 2016 11:30 AM | Permalink
Twitter: @laurenweinstein
Google+: Lauren Weinstein