On encouraging development of privacy tech, and incentives
- Tom Ritter
- 2011-11-19 @ 00:16
Posted to our blog: https://blog.crypto.is/on-privacy-preserving-technology/
This post is about a couple interesting accounts of technology,
sharing, and privacy from some unlikely sources. People don't like
labels, but from a practical standpoint it gets a bit tiresome to say
"people concerned with and interested in privacy, anonymity,
cryptography, and the like" so for short I'm just going to call us
'activists'. Don't get hung up on that, please.
First off, if you've never heard of Progressive's [Snapshot
Program - you're probably going to hate it. It's a little doo-hicky
you put in your car and monitors you. They make all sorts of
assurances and promises about how your rates won't go up, and they
don't care about how fast you drive (hah!) - but the short of it, is
it's a little monitoring device you put in your car - voluntarily.
Now obviously, there's a lot of criticism about this device from lots
of spaces but here's one I bet you didn't expect to hear it from - a
[Police-Led Intelligence](http://policeledintelligence.com/) is a "law
enforcement intelligence and crime analysis blog" written by [two
police officers](http://policeledintelligence.com/about/) and their
recent post [Snapshot: Progressive and
talks about the device.
> For those in law enforcement reading this blog, it's no news that 30
days is enough time to observe multiple illegal activities. For anyone,
it's obvious that 30 days is, in Dave's words, "Long enough to get the
full picture of their routine activity."
> Your car's computers know, for example, when you're weaving. So are you
drunk? Or are you just texting? And is texting illegal in your state?
> Speaking of terms, I like the one about where they will release data to
"comply with the law". Does this mean that I might subpoena Progressive to
obtain information on one of its customers? Pa-POW! Now that's
progressive. What a bounteous harvest might await.
> In my opinion, this may be one of those times when the cost may greatly
exceed any discount you might get.
Officer Selby makes an awful lot of the same points any of us would
make. Using my prior-defined word - something I hope we as activists
are able to do in the years ahead is work *with* companies to produce
privacy-preserving technologies that still further the state of the
art and make lives easier. No one is going to argue that electronic
toll plazas make life easier for commuters - but they also pose a
great privacy concern. Organizations and Seminars like
[PETS](http://petsymposium.org/) work to further that goal - look at
some of the talks from their last conference: "Plug-in privacy for
Smart Metering billing", "Targeted, Not Tracked: Client-side Solutions
for Privacy-Friendly Behavioral Advertising", "Privacy-friendly
Aggregation for the Smart-grid" - these aren't your usual security
conference talks. These are about *building* something - not breaking
it. I'm sure they're not perfect, but the point is, this research
isn't particularly popular but it's **critical** if we want to live in
a world where privacy is preserved. Because technology is going to
race ahead, and if we're not keeping up with it as it's developed -
our concerns aren't going to be addressed.
The second blog post is the other end of the spectrum, from Etsy's
Engineering blog - [Engineering Social
where they talk about the technical challenges of grabbing all a
facebook user's friends information and finding the best product match
> The product works by connecting with your Facebook account, analyzing
thousands or more of your friends' likes and interests, and then making
recommendations across millions of items in Etsy's marketplace. Social
commerce has been somewhat of a hot topic lately, and the gift recommender
is a social commerce feature in that it provides a new and unique shopping
experience to buy gifts for your friends and family.
Now, I actually have a lot of friends that work at Etsy, and some of
them probably helped build this - they're not bad people. But it's
awfully creepy to know that if you're on facebook - and one of your
friends can use Etsy's app - all of your likes and interests have
passed through Etsy's servers. (Possibly being stored, possibly not -
but they could have been.) Could this be done in a way that wouldn't
make me feel creeped out by Etsy? Absolutely. Did they build it that
way? I'll never know. Nor will I know if they change it. Or if they
get a National Security Letter. (Obviously I'm not actually expecting
etsy to be served a NSL and told to pull my facebook likes but... the
government has done stranger things.) Here's another area ripe for
research, development, and thought experiments.
You can sit back and dislike what the world is developing - actively
or passively - but it's not going to stop the world from innovating
and building. But you *can* dive in and work on something that
provably provides privacy *and* the features of the competitor. And
that's what I think we need more of.
tokx: it looks like you have two things going on, first is that a lot
of companies developing new technologies today aren't giving a shit
about security and that that's horrible for our privacy
tokx: and second is that it's possible for those companies to be
paying attention to our privacy
tokx: but i'm not so sure your second point is addressed as much
tokx: kind of interesting to think about because if you're talking
about the US capitalism system, people are immediately going to say,
well, what's the market incentive for companies to care about
tokx: if there is none, then there's no way you can expect them to
care about your privacy
He's right. If there is not market incentive, why would companies
care? I don't have a great answer, just two mediocre ones.
1) Some companies *do* care about doing the right thing. They're few
and far between, but they often have great brand loyalty. It's hard
to keep that up the larger you get of course.
2) Schneier repeats the mantra of increasing liability, and companies
will care about security. I think there's a secondary effect. If
they could reduce risk by just not retaining dangerous data - that'd
be another approach.