[open-government] Legal issues of linked data-different from open data?

Sun Nov 4 17:00:51 UTC 2012

Dear fellows,

the question is more complex as you will be aware of.

Even the definition of _personal data_ lead to surprisings results delivered by German studies, even 7 years ago.

Kind regards,

Gerry

Von: open-government-bounces at lists.okfn.org [mailto:open-government-bounces at lists.okfn.org] Im Auftrag von Javier Ruiz
Gesendet: Freitag, 2. November 2012 18:06
An: open-government at lists.okfn.org
Betreff: Re: [open-government] Legal issues of linked data-different from open data?

I am not an expert on the legalities of linked data either, and i am not even sure they exist yet.

I think anonymisation and its potential limitations in legal terms is mainly relevant as to whether you consider the data personal or not. If the relevant data protection regulations consider your "anonymised" dataset to be personal -- and there are very different views --  you just cannot put it online. In Europe at least you would be breaching several key data protection principles around purpose, geography and time.

You need to be satisfied that it is not personal data, even if this turns out to be wrong later on when someone links another database and starts identifying people.

I think linked data in terms of privacy raises some legal issues around who exactly has obligations to a data subject because it is processing their data: Where exactly is the data and who is doing what with it? One area to look for similarities may be cloud computing. There is research on this at Queen's Mary University.

The other issue i can think of is that you are allowing for processing of data by a third party without transferring a dataset, so I don't know what the framework would be in that case. Maybe regulations for research access to databases can be a starting point?

The cost analysis would have to be a public interest case, as privacy in Europe is becoming a proper right, and the new directive will have global reach with things like linked data.

-- 
Javier Ruiz
javier at openrightsgroup.org

+44(0)7877 911 412

@javierruiz

On Friday, 2 November 2012 at 14:29, Toby Mendel wrote:

I think we need to study carefully the issue of exactly how much risk linked data really does pose to privacy and other protected interests (like national security). Linking data is not my area of expertise but, as I said, I have not been impressed with the (admittedly few) studies I have seen on this and at least some real experts I have talked to do not believe it really is such a threat. 

Once we have a better understanding of the (real) threat, we can get into a more scientific cost benefit analysis of the sort you posit below. What are the costs of not making data available vs. the costs to privacy. I am not sure that it makes so much difference how the data is released, because I believe that technology is increasingly able to take rigid data (.pdfs) and turn it into machine readable formats. Structuring might be more of an issue, although again once you have it in machine readable formats you can start to restructure, albeit with some effort, and certainly granularity (ie how detailed and disaggregated the data is) is important.

Toby

___________________________________

Toby Mendel

Executive Director

Centre for Law and Democracy

 <mailto:toby at law-democracy.org> toby at law-democracy.org

Tel:  +1 902 431-3688

Fax: +1 902 431-3689

www.law-democracy.org

On 2 Nov 2012, at 11:19, María Täng Palma wrote:

Hello Toby,

Thank you for your answer!

That was actually my line of thought when I was reading about linked data. All the examples I found in the web explaining what linked data is always use a "person" and all the interrelations one can get about that certain person when in the web of linked data. With all the data be interlinked, cross-referenced and finally put in context, privacy is at a high risk. For another side, linked data seems to be the best way to publish PSI in a machine readable format, since linked data provides data "in a regular and well-defined structure," and the most structured the data, the easier for machines and people to re-use it. This seems to be the tendency for the future. So how will we be able to combine those things? ...

Best regards,
María

_______________________________________________

open-government mailing list

open-government at lists.okfn.org

http://lists.okfn.org/mailman/listinfo/open-government

Unsubscribe: http://lists.okfn.org/mailman/options/open-government

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.okfn.org/pipermail/open-government/attachments/20121104/e481c208/attachment-0001.html>