News Analysis

Linux Voice

Article from Issue 202/2017
Author(s):

Data protection laws are not just about keeping personal information safe; they are also about stopping the derivation of personal insights from seemingly harmless information.

Opinion

Butter, Triangulation, and Data Protection

At the end of May 2018, the new General Data Protection Regulation (GDPR) will come into effect in Europe. It creates a whole set of new responsibilities that are causing concern for businesses across the EU. It has effects outside Europe as well, because it will control the way businesses located in Europe can share data across borders, both within their company and with other companies.

While businesses are complaining about the new bureaucratic burden the GDPR creates, some privacy activists think it offers an absolute minimum level of protection in the emerging meshed society. This is not necessarily because of the way obviously confidential information is stored and used.

It seems obvious why we should be concerned about big chunks of personal data, but why should we care about protecting small details such as our date of birth, parents' names, postal code, and so on? Why does it matter when we're asked for them by someone with no need to know them?

The reason is, personal info can be used for triangulation to breathe meaning into otherwise anonymous data. Since they are likely to appear as a field in every database table that records information about us, they can readily be used as shared keys to allow data manipulation. What does that mean? Here's a (currently completely fictional) example to explain.

You are at the checkout in your local supermarket. They scan the pack of butter you want to buy along with the rest of your shopping, and then at the end of the packing as you prepare to pay, you hand over your supermarket loyalty card. The assistant looks at the screen, then reaches for the voucher printer, and pulls a form from it. He places it on the counter and gives you a pen. "Here, sign this." You look at it in surprise. It is a liability waiver, with your name at the top. The text simply says "I absolve the store of all consequences for my purchase of butter."

How did this happen? The store doesn't know your health status; they just know it's in their interest to get that waiver signed. Something like this happened:

  • Their insurance company has used your name, postal code, and year of birth – effectively a unique identifier – as a "shared key" to identify you.
  • They have used this shared key to ask holders of health records, past purchases at other stores, data from advertising tracking on web sites, and other information about you to each give a classification (without actually sharing the raw data), and they received back data about the data ("metadata") that they've then kept in a new database.
  • At the time of purchase, they have fed this metadata into an actuarial model to see if anything in your basket triggers any indication of elevated risk.
  • Because it did, and because people with similar profiles have sought class actions against contributors to health issues, they flagged you to this store as a litigation risk.
  • The store gets a discount on their liability insurance if they get waivers from all at-risk customers, hence the waiver form.
  • The store also gets a reduction on their insurance if they will provide metadata about your loyalty card purchases.

This is triangulation. No individual data item discloses private information you really care about, but gathered together, it can be deduced and used without consent. Triangulation is how anonymous data can be de-anonymized, and why assurances that your data has been anonymized and is thus safe to share should be treated with skepticism. The principle of least privilege (check it out on Wikipedia!) should inform us everywhere in our lives and is why we should support data protection laws like the EU's new GDPR.

The Author

Simon Phipps is a board member of the Open Source Initiative, the Open Rights Group, and The Document Foundation (makers of LibreOffice).

Buy this article as PDF

Express-Checkout as PDF
Price $2.95
(incl. VAT)

Buy Linux Magazine

SINGLE ISSUES
 
SUBSCRIPTIONS
 
TABLET & SMARTPHONE APPS
Get it on Google Play

US / Canada

Get it on Google Play

UK / Australia

Related content

  • News

    You don't control that device you think you own.

  • Data Security in AWS

    As a cloud market leader, Amazon Web Services has had to  put a great deal of thought into data security. Encryption options and key management play an important role.

  • Browser Cleanup

    Web browsers collect a large amount of data about the user’s browsing habits. If you care about privacy, you might want to clean up your browser and configure some custom privacy settings.

  • Data Transfer Project

    The Data Transfer Project wants to make it easier to move your data between social media sites.

  • Clonezilla: Clone for Windows and Linux Machines

    The latest version of the Norton Ghost clone for Linux brings improved support for Windows clients.

comments powered by Disqus
Subscribe to our Linux Newsletters
Find Linux and Open Source Jobs
Subscribe to our ADMIN Newsletters

Support Our Work

Linux Magazine content is made possible with support from readers like you. Please consider contributing when you’ve found an article to be beneficial.

Learn More

News