The Downsides of Google's Chrome Security Push

https://lauren.vortex.com/2016/09/08/the-downsides-of-googles-chrome-security-push


Google has world class security and privacy teams, but I continue to
have misgivings about certain aspects of their Chrome browser security
push -- particularly regarding warnings to users when connections are
using unencrypted http: as opposed to https: encryption.

While the push to encrypt Internet connections by default is a
laudable one, it is also essential that fundamental aspects of
practicality and user reactions also be carefully considered.

I touched on some of this over a year ago in "Falling Into the
Encryption Trap" ( https://lauren.vortex.com/archive/001108.html ) --
but now that Google has made more explicit their plans for browser
address bar warnings to users regarding http:  connections, I'm again
concerned.

Apparently in January of next year Google intends to replace the
current quite reasonable "information circle" indicating non-encrypted
pages, with an explicit "Not secure" warning -- ultimately to be
displayed in bright red with a danger triangle.

I am absolutely certain -- based on the many queries I receive
routinely from users who are already confused and concerned about
other security warnings they see and misunderstand -- that the
escalation to these sorts of warnings by Chrome will vastly and
unnecessarily increase confusion and even panic among significant
categories of non-techie users when accessing various sites important
to them.

Because the truth of the matter is that it remains both impractical
and unnecessary for all sites to convert to https: at this time.

It is certainly true that theoretically any site could become a vector
for misinformation or malware via man-in-the-middle manipulation of
their connections, and the use of various insecure and/or poorly
managed ad networks increases the risks in this context.

But as a practical matter, the vast majority of exploits that users
must contend with do not come from the manipulation of Internet
connections. Rather, infections via email phishing, contaminated
sites, and similar techniques represent the overwhelming majority of
successful attack vectors.

Still, it is inarguable that all else being equal, having all
connections as encrypted https: rather than unencrypted http: is
extremely desirable.

Unfortunately, all else isn't equal.

There are uncountably vast numbers of legacy sites that provide widely
referenced information to enormous numbers of users, yet do not sell
anything, don't collect usernames or passwords or other private
information, and don't participate in any ad networks.

Many of these sites have been online not just for many years, but even
for decades. They typically use older software systems that are
difficult or impractical to directly update, and frequently operate on
a shoestring (or even zero) budget, while not creating any income at
all.

It will frequently prove impossible from a money and/or time
standpoint for the operators of such sites to convert to https: -- yet
Chrome's warning system will likely confuse their users into assuming
that they are actually being spied on -- rather than the actual fact
that such surveillance is in any given case theoretical (and in
practice an extremely low probability) on those individual
connections.

And while the cost of encryption certificates has now dropped to zero
with the advent of services such as "Let's Encrypt" -- the effort
required to actually make them work can be anything but trivial.

I recently converted all of my sites, some of very long standing, to
https: using Let's Encrypt. Even though my sites are not fancy in any
way, it was an enormous amount of work, and required every ounce of
knowledge I had regarding the sites' internal architectures. While
Let's Encrypt promotes scripts to supposedly handle such conversions
automatically, I cannot recommend those procedures except for the very
most trivial and simplistic of sites -- anything beyond that and
you're likely to end up with a mangled site configuration nightmare --
you'd better have good backups handy!

I'm frankly uncertain how to best achieve a practical compromise
position regarding browser security warnings.

I do know that a scary red "Not secure" warning is likely to
unnecessarily panic many users and unreasonably disadvantage many
sites.

This is especially true when there is no explicit indication to users
as to how they can obtain more information about that warning -- such
as what does it really mean in terms of actual risks? -- in language
that non-techies are likely to actually understand. Even now, the
security details that Chrome provides if one knows to click on the
address bar security icon are pretty much technical gobbledygook as
far as most users are concerned.

My sense is that despite their great skills in privacy and security
matters, Google has not genuinely considered the impacts of their
upcoming browser warnings on significant segments of the user and site
populations, who by and large do not live 24/7 in the same rarefied
security worlds as do many of us.

Luckily, this is a fixable problem, if Google is willing to put forth
the effort and outreach to fix it. I respectively urge them to do so.

Be seeing you.

--Lauren--
Care About Science and Tech? Our Job One: STOP TRUMP: 
https://vortex.com/stop-trump
 - - -
Lauren Weinstein ([email protected]): https://www.vortex.com/lauren 
Lauren's Blog: https://lauren.vortex.com
Founder: Network Neutrality Squad: https://www.nnsquad.org 
         PRIVACY Forum: https://www.vortex.com/privacy-info
Co-Founder: People For Internet Responsibility: https://www.pfir.org/pfir-info
Member: ACM Committee on Computers and Public Policy
Google+: https://google.com/+LaurenWeinstein
Twitter: https://twitter.com/laurenweinstein
Tel: +1 (818) 225-2800
I have consulted to Google, but I am not currently 
doing so -- my opinions expressed here are mine alone.
 - - -
The correct term is "Internet" NOT "internet" -- please don't 
fall into the trap of using the latter. It's just plain wrong!
_______________________________________________
pfir mailing list
https://lists.pfir.org/mailman/listinfo/pfir

Reply via email to