Re: [projectvrm] De identification of data

[Adrian Gropper < >]

De-identifying data is just a face-saving way of selling that which you don't own. Personal data has value and that value might or might not be shared with the service provider that helped create the personal data. Even so, the sharing still needs consent and, well, sharing.

For the most part, the sale of de-identified data benefits only the service provider and de-identification is merely a tactic to avoid asking the person for consent. This is most evident in US healthcare where prices are opaque and quality measures non-existent. Healthcare supports a growing market for de-identified data that may well exceed $20 B at the same time that US consumers are being overcharged by $1 T compared to 10 other developed economies. Does anyone really think these two numbers are unrelated?

Adrian

On Monday, July 14, 2014, John Wunderlich < "> > wrote:

If you approach de-identification as a binary set, where it is a failure if even one person out of a million might theoretically be re-identified, the argument might make sense. In a real world risk management scenario, it fails to be a reasonable or sustainable argument.

Latest entry is here:

http://www.innovationfiles.org/meet-the-new-de-identification-deniers/

The debate is fairly heated. Full disclosure, I have worked for Ann Cavoukian and with Khaled El Emam.

John Wunderlich

Privacist @PrivacyCDN

On 9 July 2014 20:28, John Havens < ');" target="_blank"> > wrote:

Found this very intriguing and wanted to get the group's opinion:

http://boingboing.net/2014/07/09/big-data-should-not-be-a-faith.html?utm_content=buffere752f&utm_medium=social&utm_source=twitter.com&utm_campaign=buffer

Thanks!

John Havens

-- 
Adrian Gropper MD