You would think
that all those tinfoil hats, you wouldn't actually want people working on cures for cancer using big data..
Google’s DeepMind operation insists UK patients have nothing to worry about now that Google has absorbed the subsidiary - but lawyers and privacy campaigners have raised doubts. DeepMind told The Reg: “It is false to say that Google is 'absorbing' data. This data is not DeepMind’s or Google’s – it belongs to our partners, …
All those Google shills, you'd almost think they wouldn't want anyone to scrutinise a massive advertising company potentially getting access to medical records when that company happens to make almost all of its money by mercilessly extracting personal information from as many people as it possibly can...
you wouldn't actually want people working on cures for cancer using big data
A Cure for Cancer is a novel by British fantasy and science fiction writer Michael Moorcock, first published in London 1971 by Allison and Busby. The book is part of Moorcock's long-running Jerry Cornelius series.
The second novel of the sequence is essentially a collage of absurdist vignettes, many of which first appeared in an eclectic range of British and American magazines.
Jerry inhabits a world at war with itself and, armed only with an occasional "vibragun" appears to fight "against history" for the freedom of "randomness" against the straitlaced conventions exemplified by his brother Frank. In the end Jerry's quest, oblique as it is, is perhaps more artistic than political.
Sounds rather apposite.
"It's clearly mentioned this is anonymous data"
You say that as if it's meaningful. Just calling the data "anonymized" is meaningless. Most popular methods of "anonymizing" data is ineffective in the face of Big Data, which often makes deanonymizing data very easy to do. Some forms can be effective (such as only keeping data aggregated with that of large numbers of other people and deleting individual data records), but we don't know if that's what they're doing or not.
If not, then "anonymous data" is effectively a lie.
They produced an analysis for the Royal Free, see https://s3-eu-west-1.amazonaws.com/files.royalfree.nhs.uk/Reporting/Streams_Report.pdf.
I have read the foreword, the management summary and the bit tacked on about GDPR.
Then I got bored.
It would seem that any use of patient personal data for purposes other than those explicitly set out in the contract will be illegal, either by DeepThroat or Google.
Doesn’t mean they won’t do it though.
Also, should that AWS s3 bucket have been open?
"It would seem that any use of patient personal data for purposes other than those explicitly set out in the contract will be illegal"
This is the Linklaters report paid for by the Royal Free. In it, everything's lovely. There's nothing to worry about. Unicorns have even been sighted in the meadow.
Story: Audit of DeepMind deal with NHS trust: It checks out, nothing to see here
No one else seems particularly reassured. The only opinion that matters here though is the ICO and they have yet to deliver their verdict.
Do you feel more confident now you know that Google/DeepMind doesn't actually *know* who owns the data, and falsely asserts that the Trust owns it?
I don’t get this. Surely you would just look at the contract the Royal Free has with DeepMind (released under FoIA).
It says the patient records can only be used to provide the App, must be stored within England, not disclosed to any third party (which includes Google), cannot be linked to any other data etc.
I just don’t see anyway DeepMind could provide any access to this data to the rest of Google without breaching that contract. That’s just a fact.
Biting the hand that feeds IT © 1998–2018