Celebrated eye hospital Moorfields lets Google eyeball 1 million scans Famous eye hospital Moorfields has agreed to give Google’s DeepMind access to one million anonymous eye scans as a part of a machine learning study intended to spot early signs of sight loss. Explicit patient consent is not required because the scans are historic, meaning the results won’t affect the care of current patients. …
Correlating with scan results to Android "id by face"? Google has more than enough information to de-anonymize it.
Hmm...
Revelation 13:15-18
The second beast was given power to give breath to the image of the first beast, so that the image could speak and cause all who refused to worship the image to be killed. It also forced all people, great and small, rich and poor, free and slave, to receive a mark on their right hands or on their foreheads, so that they could not buy or sell unless they had the mark, which is the name of the beast or the number of its name.
Here's the thing about Revelation. It's so fantastically specific on details which are relevant across all of history (eg the importance of the ability to buy and sell), while being completely vague on the details which could narrow down the time period which it's talking about, that it has been interpreted as applying to everything from Nero, through Napoleon, to Hitler, and now to all sorts of body implant type things.
It was written to offer encouragement under persecution, no matter when in history that persecution happens, and it's been astonishingly well written for that. Google isn't, really, persecuting. So I'd go out on a limb, and say that, perhaps, it's not really relevant here.
I know it won't make me popular to say this here, but in pure Biblical terms, Google probably aren't the Antichrist.
Re: 'shrooms' - that's usually a convenient explanation for anyone who can't be bothered to find out what 'Revelations', or the family of literature it belongs to - 'apocalyptic', are actually all about.
So, no - the writer almost certainly wasn't 'on the shrooms', but knew exactly what he was on about. Even if we struggle to comprehend it today, in it's day the imagery and allusions would have made perfectly good sense to anyone educated in the 'language'.
For a modern example, although generally far less lurid, we only need to look at some of the writings of people under violently oppressive governments to see how imagery and 'code words' - perfectly innocent to the uninformed reader - are used to communicate effectively. Apocalyptic is another style all together, but the intent is largely the same.
‘Did you know, young lady,’ said Watkin to her, ‘that the Book of Revelation was written on Patmos? It was indeed. By Saint John the Divine, as you know. To me it shows very clear signs of having been written while waiting for a ferry. Oh, yes, I think so. It starts off, doesn’t it, with that kind of dreaminess you get when you’re killing time, getting bored, you know, just making things up, and then gradually grows to a sort of climax of hallucinatory despair. I find that very suggestive. Perhaps you should write a paper on it.’ He nodded at her.
Douglas Adams - Dirk Gently's Holistic Detective Agency
There has been, of late, a shift in The Register's attitude towards data collection in the direction of sheer bloody paranoia.
If by giving Google this data we can save the sight of just one person through earlier diagnosis then it's worth it, because the cost to the people whose data is made available is zero. Zilch. Nada.
If you're going to write articles like this I trust that you don't use a smartphone; you don't browse the internet outside of The Reg; you don't have a smart meter; you don't drive; you don't use credit cards; you don't buy from big shops; you don't have WiFi at home; your music collection is on CD or vinyl; your books are paper; etc etc.
Bear in mind that this policy of evaluating treatment methods by careful analysis of big data is becoming the norm because it WORKS. Yes, precautions need to be taken, and yes, if you're truly paranoid you could assume that your phone is trying to get a closer look at your eyes to see if you're on the list, but there are bigger things in the world to be worried about. Theresa May, for example.
have a down vote. as noted above google is being given access to biometric identifiers. enslaving an entire society to save one persons sight is a crap deal. this is the "won't someone think of the children" argument writ large.
OK OK, I'm exaggerating. somewhat.
I also wonder whether the data is actually anonymous, or would it happen to have the NHS number tacked on as well? neither google or the NHS are to be trusted here.
If I could put this paranoia in context here: If "Google + Deepmind + 1 million anonymous eyeballs = Oh Noes!", what does a far more advanced and complete AI + *300* million healthcare recordsets mean? Note how this article has only has 5 comments:
http://www.theregister.co.uk/2016/02/18/ibm_watson_in_26bn_grab_to_completely_own_300m_patient_lives_data/
Deepmind played a boardgame. Watson played Jeopardy, live on TV, absolutely crushing the human competition. Fear Google if you want..
This post has been deleted by its author
"Yes, precautions need to be taken"
This is the crux of the matter. Have adequate precautions been taken? Was patient consent asked for, let alone given? Is there sufficient information to prevent Google de-anonymising the data? What legal steps and real checks exist to ensure they don't try?
The fact that it might benefit some patients is not at issue but it's not a valid excuse for not strictly adhering to the requirements of the DPA. "It's for your own good" is exactly the justification used by the Mays of this world. Legitimising it by means of benign medical research is not a good idea.
That's exactly what I did. There is no way of opting out of this set out in the Q&A. So I sent them a section 10 DPA notice.
It's a shame I'm having to take time out of my day to send legal threats to the NHS when a simple "we would like to use your personal data for this project, would you object" would have been fine by me.
And does the UK get a free pass on any IP related to this, such as no having to pay for any diagnostic tests, procedures or therapies that result from the use of this data? Let alone non medical use such as new biometric retina analysis for identification or other non medical related discoveries derived from the use of UK citizens data?
"Famous eye hospital Moorfields has agreed to give Google’s DeepMind access to one million anonymous eye scans as a part of a machine learning study intended to spot early signs of sight loss members of the Resistance before they attempt to destroy the foundations of Skynet."
There, FTFY.
Heh, true :)
What is the worst case scenario here anyway? That they'll be able to use retinal scans to identify people they've already scanned the retinas of?
Has this $UNSPECIFIED_EVIL we're all frightened of actually been defined anywhere? Is there any danger at all that we are, in fact, all looking rather silly?
we fucked up. We didn't ask you for the opinion, because first, we had no clue how important it is, and when we realized, we shrugged our shoulders and made quarter-hearted attempts to make it look like it's nothing serious.
This is being "more open".
However, in reality, being "more open" is likely to be:
"As we fear bad publicity, we will make it look like it's not us guv and spout the usual bullshit about how we ALWAYS care about patients records and their privacy and ALWAYS take it extremely seriously and it is ALWAYS our top priority, blah blah blah.
Ah, and I bet they'lll say something how this could save countless patients' lives, etc.
I agree with the sentiment here. There are some things which help medicine that people actively opt in to. Examples that come to mind are blood donation, organ donation, entry into drug trials.
But there are also examples where sharing patient info can help others. Real world example, from a member of my close family....exhibited a weird set of symptoms that were potentially dangerous (heart-related). Much medical investigation later, the medics found the cause of the problem and it was something that they'd never seen before but it was recognised could happen in other people.
Within months, the case was written up and being talked about in The Lancet and elsewhere. No consultation with the patient about whether or not this was going to happen - it happened because it could help advance medicine and help other people.
@ Uberseehandel
Medical research is not incompatible with following proper data protection procedure. Has this been done here? If it has, all well and good but questions about successful anonymisation have been around for a good while, even when the likes of Google weren't involved. It isn't unreasonable that such a project should receive careful scrutiny along these lines.
Actually this is nothing to do with wht we are talking about.
Google have ZERO interest in saving peoples eyesite, they are looking for ways to make money from data (generally personal, but whatever).
They get that data given to them by providing a service, like gmail, or by offering some service to the organisation in question. In this case the NHS
This data is then used whoever they can to make a profit.
Off the top of my head I can see uses in working out what heredatory illnesses a person may suffer from purly from a scan of their eyes, and as mentioned before, its impossible to anonymise data thats unique to a person ESPECIALLY when the organisation in queston already scrapes that informatiion (or enough information) to link high resolution scans to a specific person.
This is why we are supposedly given some protection in the "Data Protection Act" from OUR information being given out without our consent. And THIS is why people on this thread are angry.
Yes this Act is relatively useless, because, obviosuly, once the data is out its never going to disappear, the horse has bolted, this does not mean that we should ignore the people who opened the gates, but I doubt anything will come of this as the NHS/Police/Local Government are actualy above any laws that the common people are bound by
Interesting... so progress at all costs? Much of the data from the WWII concentration camps medical "tests" (I use the word broadly here) will not be used by ethical institutions. Is that a good thing or a bad? Think about the German hypothermia tests, for example which is the basis for much of the knowledge of hypothrermia. Or the so-called "twin" tests were one twin was given something and observed. When the test was over, both were killed and autopsied. Is these good or bad? Is the methods of gathering the data good or bad?
What's going on here may not be a bad thing. What's being argued is the strings attached to it. There's a feeling amongst us that not all is Kosher with this deal.
[IAAD]
If I were a patient then I would be OK with this but I would not be OK with what Royal Free did which was to release the whole lot of identifiable patient information which is totally unacceptable. If Royal Free were a company then the Board would have to resign and there would be an eye-watering fine.
How about a rule that if there is a hospital data breach then the directors should pay the fine themselves AND have their medical data published on wikileaks?
Just advertising companies?
If you took your sentence above and replaced "advertising company" with "for-profit company engaged in <business>" would that make it seem any more or less OK?
I reckon it'd still fall on the "not OK" side of your fence, as it'd be seen as a way of a company making money - directly or indirectly - from other peoples' data
It's not so much an advertising company, but a company that monetises people's personal information. Whether that's tailoring ads, or flogging data to insurers. Yes, there may be some medical benefits, but also lots of potential commercial benefits to Google. Especially if they get more of these deals and start de-anoymising and cross-linking. Data controllers could try putting limits on use and sharing, but once the data is out there, they've lost control of it.
It bothers me a deeply for-profit company with no incentive to do anything "for the public good" is being given this data.
It's easy to foresee them hitting huge conflict of interest issues if they keep developing their medical tech and applying it to large population groups without consent (eg NHS data).
For example, lets say they apply their (further developed) screening for some kind of nasty disease (no idea what, it's just an example). Lets say something that's likely to be fatal in under 5 years. Lets say they screen the data for all of the NHS patients they have access to.
When they get positive diagnosis results, the choice goes like this:
a) Will this person be able to influence things to Google's benefit or detriment? (eg are they involved in any investigations against Google, or likely to be)
b) Are they a child or close relative of someone who is?
For the example, suppose it's someone who's investigating them for serious stuff, with potential billion $ fines. (eg senior EU Commission staff). It wouldn't be surprising (to me) if the notification gets "lost in the mail".
The implications of who to notify, how to notify them, and when, can be substantially ethically challenging. It's not something a Google company should be trusted with (my opinion).
What is really sad is that out of the 5 leading cases of blindness:
1) Diabetic Retinopathy - A sympton of diabetes which is for the majority of people a result of bad diet.
2) Refractive Error - Can be corrected to a degree by optician but extreme cases exist.
3) Cataracts - Has a very successful success rate with surgery and most patients will have excellent QOL. Though you'll likely lose your reading ability without glasses.
4) Glaucoma - Is somewhat manageable if found early as it's down to high eye pressure.
5) Macular Degeneration - This is the bad one. It depends on whether it is dry or wet (dry is bad, wet is worse), but currently this can only be managed but will typically only get worse over time. It robs you of your central vision (ie your macular).
So out of the five, four of them are mostly manageable with modern medicine or diet changes. The fifth one hopefully will see some scientific breakthrough in the next few decades.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
