[MyData & Open Data] Economic signalling, open data, my data, and privacy unraveling

Thu Apr 17 12:24:40 UTC 2014

I struggle too, however I think that is simply a function of where we are
at the moment.  We've only mapped the tip of the data/privacy iceberg.
 Here's another example [1] (courtesy of Bruce Schneier) that points out
that opting out of geo-location on Twitter doesn't mean that your tweets
can't be used to geo-locate you.  Apparently the similarity of your tweets
to those who do choose to geo-locate their tweets is enough to identify a
user's city for 58% of users.  Not exactly earth shattering statistics but
one could imagine how this might be combined with other public data to
become much more granular.

Goes back to points made earlier about the London cycling data, being used
with other geo-datasets posing significant risks to re-identification . As
rightly suspected, using customer IDs was a 'mistake' which has since been
rectified by TFL (see
here<http://skift.com/2014/04/16/london-transports-bike-share-privacy-slip-raises-concerns/>
).

Re. the point about Other people's privacy choices affect ours.

I wonder though if there is opportunity to debate further how how
individuals choosing to open 'mydata', could affect the privacy of other
person's who's identity is linked somehow  to this individual, and if there
are any safeguards against this?

 Perhaps the answer lies in the mathematics of differential privacy[2][3]
where noise is added to each data query in order to protect privacy.  I
confess I don't fully understand differential privacy yet but it does seem
like we need some more powerful combination of regulation and tools like
differential privacy in order to protect our right to privacy.  This may
offer some hope for the intersection of personal data and open data.

Perhaps it is. Differential privacy was also well discussed in this
recent document
<http://ec.europa.eu/justice/data-protection/article-29/documentation/opinion-recommendation/files/2014/wp216_en.pdf>on
the Opinions on Anonymisation Techniques. Worth a good read.

Sally

On 16 April 2014 19:30, Song, Stephen <stephen.song at gmail.com> wrote:

> Hi Sally,
>
> On 11 April 2014 12:33, Sally Deffor <sally.deffor at okfn.org> wrote:
>
>> He makes strong arguments and as I struggle to see what sort of case any
>> regulatory framework could possible make to counter individual signalling,
>> I would imagine that what he proposes as a solution to 'unchecked
>> signalling' is beyond the realm of informational privacy advocates? In
>> practice, individuals tend to weigh personal considerations against what is
>> the acceptable norm when choosing to fully disclose depending on if they
>> hope to economically or not.
>>
>
> I struggle too, however I think that is simply a function of where we are
> at the moment.  We've only mapped the tip of the data/privacy iceberg.
>  Here's another example [1] (courtesy of Bruce Schneier) that points out
> that opting out of geo-location on Twitter doesn't mean that your tweets
> can't be used to geo-locate you.  Apparently the similarity of your tweets
> to those who do choose to geo-locate their tweets is enough to identify a
> user's city for 58% of users.  Not exactly earth shattering statistics but
> one could imagine how this might be combined with other public data to
> become much more granular.  Other people's privacy choices affect ours.
>
> In terms "acceptable norms", I think 99% of people have absolutely no clue
> as to how much privacy they have already given up through the general use
> of technology whether through the Internet or a turnstile.  Further I think
> we have only scratched the surface of how data can be jig-sawed together to
> reveal even more personal information.
>
> Perhaps the answer lies in the mathematics of differential privacy[2][3]
> where noise is added to each data query in order to protect privacy.  I
> confess I don't fully understand differential privacy yet but it does seem
> like we need some more powerful combination of regulation and tools like
> differential privacy in order to protect our right to privacy.  This may
> offer some hope for the intersection of personal data and open data.
>
> Regards... Steve
>
> [1]
> http://arstechnica.com/business/2014/03/lack-of-twitter-geotags-cant-stop-researchers-from-getting-location/ (research paper at
> http://arxiv.org/pdf/1403.2345v1.pdf)
> [2]  https://en.wikipedia.org/wiki/Differential_privacy
> [3]  Differentially Private Publication of Sparse Data -
> http://arxiv.org/abs/1103.0825
>
>
>> On 11 April 2014 14:44, Song, Stephen <stephen.song at gmail.com> wrote:
>>
>>> Hi all,
>>>
>>> I'd like to raise another issue here related to the role of individual
>>> choice in making data public.  In this paper by Scott Peppet
>>>
>>> Unraveling Privacy: The Personal Prospectus & the Threat of a Full
>>> Disclosure Future
>>> http://www.scottpeppet.com/2012/04/unraveling-privacy/
>>>
>>> Peppet argues that by choosing to make our data public, we are forcing
>>> an implied choice on others who don't, about which people will make
>>> assumptions.  He gives the example of an orange seller.
>>>
>>> "The classic example of unraveling imagines a buyer inspecting a  crate
>>> of oranges.  The quantity of oranges in the crate is unknown and opening
>>> the crate before purchase is unwise because the oranges will rot before
>>> transport. There are stiff penalties for lying, but no duty on the part of
>>> the seller to disclose the number of oranges in the crate. The number of
>>> oranges will be easy to verify once the crate is delivered and opened. The
>>> buyer believes that there can’t be more than one hundred oranges.
>>>
>>> The unraveling effect posits that all sellers will fully disclose the
>>> number of oranges in the crate, regardless of how many their crate
>>> contains. Begin with the choice faced by a seller with one hundred oranges
>>> in his crate. If the seller stays silent, the buyer will assume there are
>>> fewer than one hundred oranges and will be unwilling to pay for the full
>>> amount. The seller with one hundred oranges will therefore disclose and
>>> charge full price. Now consider the choice of a seller with ninety nine
>>> oranges. If this seller stays quiet, the buyer will assume that there are
>>> fewer than ninety nine oranges and will discount accordingly. The silent
>>> seller gets pooled with all the lower-value sellers, to his disadvantage.
>>> He will therefore disclose.  And so it goes, until one reaches the seller
>>> with only one orange and the unraveling is complete."
>>>
>>> You can imagine this applied to health data, fitness data, driving data,
>>> etc.  He sums it up by saying:
>>>
>>> "In a signaling economy, even if individuals have control over their
>>> personal information, that control is itself the undoing of their privacy"
>>>
>>> The paper left more questions for me than it answered.  How significant
>>> a factor is signalling?  How can it be weighed against individual freedom
>>> of choice?  This sort of thing already happens with insurance companies.
>>>  Is the risk the same in other realms of open data?  His argument strikes
>>> me as convincing but theoretical, a possible extrapolation of things as
>>> they are.  I'd love to hear other opinions on this paper.
>>>
>>> Regards... Steve
>>>
>>>
>>> _______________________________________________
>>> mydata-open-data mailing list
>>> mydata-open-data at lists.okfn.org
>>> https://lists.okfn.org/mailman/listinfo/mydata-open-data
>>>
>>>
>>
>>
>> --
>>
>>
>> *Sally DefforOpen Data & Privacy Project Coordinator |
>> skype:deffor.selase | @SDeffor | +44 (0)7774 734206
>> <%2B44%20%280%297774%20734206> The **Open Knowledge Foundation*<http://okfn.org/>
>>
>> *Empowering through Open Knowledge**http://www.okfn.org*<http://www.okfn.org/>*
>> | **@okfn* <https://twitter.com/OKFN>* | **OKF on Facebook*<http://www.facebook.com/OKFNetwork>*
>> | **Blog* <http://blog.okfn.org/>* | **Newsletter*<http://okfn.org/?s=Newsletter>
>>
>
>
>
> --
> Steve Song
> +1 902 529 0046
> http://manypossibilities.net
> http://villagetelco.org
>
> _______________________________________________
> mydata-open-data mailing list
> mydata-open-data at lists.okfn.org
> https://lists.okfn.org/mailman/listinfo/mydata-open-data
>
>

-- 

*Sally DefforOpen Data & Privacy Project Coordinator | skype:deffor.selase
| @SDeffor | +44 (0)7774 734206 The **Open Knowledge
Foundation*<http://okfn.org/>

*Empowering through Open Knowledge**http://www.okfn.org*<http://www.okfn.org/>*
| **@okfn* <https://twitter.com/OKFN>* | **OKF on
Facebook*<http://www.facebook.com/OKFNetwork>*
| **Blog* <http://blog.okfn.org/>* |
**Newsletter*<http://okfn.org/?s=Newsletter>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/mydata-open-data/attachments/20140417/cdcc7df2/attachment-0003.html>