topic Re-identification easier than was thought in Privacy

Re-identification easier than was thought

Steve-Wilme — Mon, 29 Jul 2019 09:53:09 GMT

Next time you think you've anonymised that big data set by removing obvious identifiers, look more closely:

https://www.theguardian.com/technology/2019/jul/23/anonymised-data-never-be-anonymous-enough-study-finds

Re: Re-identification easier than was thought

AppDefects — Mon, 29 Jul 2019 13:40:32 GMT

@Steve-Wilme wrote:
https://www.theguardian.com/technology/2019/jul/23/anonymised-data-never-be-anonymous-enough-study-finds

I'm glad to see that the original Working Party 29 Anonymisation Techniques are getting a fresh new research perspective. How data is shared today between partner organizations and data brokers is totally broken. Recently, Google has tried to make its implementation of Private Join and Compute publicly accessible through a GitHub repo but it remains to be seen who is adopting the techniques...

Re: Re-identification easier than was thought

Steve-Wilme — Mon, 29 Jul 2019 14:36:40 GMT

The UKAN decision making framework is a useful resource in this area:

https://ukanon.net/ukan-resources/ukan-decision-making-framework/

I've yet to find a developer or big data analytics person who's actually read it though 😞

Re: Re-identification easier than was thought

AppDefects — Mon, 29 Jul 2019 16:46:33 GMT

@Steve-Wilme wrote:
The UKAN decision making framework is a useful resource in this area:
https://ukanon.net/ukan-resources/ukan-decision-making-framework/

I've yet to find a developer or big data analytics person who's actually read it though 😞

You just found a CSSLP that has read and preaches this guidance! What I do differently though to complement the framework is teach how-to implement anonymisation algorithms, design UI's that are privacy-preserving. and design data sharing mechanisms. That is what developers and statisticians care about. Implementing code!