The video above captures the touch in and touch out of Oyster cards across London over a single day. Think about it. The ebb and flow of 7 million Oyster cards; a day in the life of the capital.
It’s pretty fancy eye candy from ‘big data’. But suppose you had to choose between big data or social media. Between control and anarchy. Between the NSA and Edward Snowden – which would you go for?
It’s a false dichotomy, of course, since we don’t have to choose. Or do we? As the digital revolution continues to chomp its way through the body politic the underlying patterns become clearer: big data creates new concentrations of power whilst voice and social media disperse old ones.
Consequently, each appeals to a different constituency. Are the patterns of power and capital associated with the two megatrends of big data and social media so different as to drive different parts of the system in consistently different directions?
Pulling power apart
Unsurprisingly, elites are entranced by big data. Suddenly, the meta-patterns are made manifest, and the needle is glinting in the haystack – ready to be seized by the all seeing eye at GCHQ (or NHS England).
The data is splurging everywhere, the fire hydrants of information roaring at full throttle. Why would anyone charged with preventing the next terrorist attack (or epidemic) settle for anything less?
By the same token, social media is tailor-made for citizens. Put £500 of digital capital (aka a smart phone) in everybody’s hands and the world turns webby: instant knowledge, total mapping, uber-transparency, democratised voice, citizen journalism, do it yourself pornography.
It’s just what the doctor ordered when you want to campaign against a NICE protocol or shame Virgin Trains for being late again. The world in which everybody has a public voice smooths out the high points of control, flattens the power differentials, and generally makes being in charge a lot less fun than it used to be.
So, unsurprisingly, it gives elites the heebee jeebies. It makes them cast their eyes back lovingly to the 20th Century, when they controlled more stuff, people were deferential, and hierarchies had the world at their feet.
Populist talk, centralising tendencies
This dichotomy is often not made clear, because big data advocates like to talk as if they are part of the social media world. And, at first glance, the ideology down at the big data store looks almost egalitarian.
The state is clearly useless at liberating the value tied up in its massive databases – so let’s free the stuff so everyone can mash it up. This reversal of the state’s long history of holding its own closely to its breast is as welcome as it is surprising. But it turns out that big data comes with deeply centralising DNA.
Most obviously power naturally flows to those who own the platforms (Facebook, Google, Amazon and so forth) or those who own the pipes (GCHQ).
Jaron Lanier in his recent book, ‘Who owns the future?’ calls these the ‘siren servers’ – the points of uber-leverage in the system. Unsurprisingly, you don’t find Facebook or government ‘freeing’ these.
Secondly, the power of denominators is immense. Knowing that your trust discharged 202 people on the right drugs is data. Knowing whether this is 98% or 25% of the relevant cohort is information.
Just look at care.data – the programme that NHS England and the Health and Social Care Information Centre are promoting to expand the Hospital Episode Statistics and link them to new data sources.
You can see that utility of the programme rises rapidly as it approaches 100% coverage. It’s just so much better to have the whole data set.
This may explain why care.data seems biased against people opting out; the interpretation that Nature and other commenters have put on the public information leaflet’s failure to include an ‘opt out’ form.
However, the need to define denominators forces big data to get bigger, handing power back to organisations big enough to handle it.
Finally, actually using data in the real world still depends on the old forms of human-mediated control like due process and information governance. But if you are charged with stopping terrorist attacks, or fighting the next epidemic, due process is far too slow and information governance a complete pain.
Far better to agree any governance centrally and then bake it into the code and the API access. That way everything can move at the speed of light.
And, as we know from Facebook, it’s a cinch technically to tweak privacy settings up and down once they are embedded in the platform.
A sign of things to come
The cognitive dissonance inherent between social media and big data is being played out around care.data.
For those developing care.data it is self-evident that it’s fine to suck in lots of identifiable data, because it will only be pumped out to deserving researchers, once it has been ‘pseudo-anonymised’, and this process will be secure.
For patients and citizens, ‘security’ is more problematic. They see the Snowden revelations and wonder where all that data will go in the future.
They remember just how often officialdom has left memory sticks in railway carriages and they can see their own inboxes filling with phishing spam and hacked Twitter messages. This is the process by which social media eats away at big data’s narrative.
At the moment, care.data seems to be winning the battle for hearts and minds. In part this is because people view ‘the NHS’ more benignly than ‘the government’. In part it is because care.data doesn’t try very hard to explain the possible down sides of care.data.
The leaflet that went to 22 million households failed to include anything remotely like a ‘contra-indications and side effects’ section.
Taking a wider view it is telling that a proposed EU directive may stymie the whole venture, by insisting that citizens must always give informed consent to their personal data being used for any purpose.
Leaving any ‘Bah, Euro humbug!’ reaction aside, there is clearly a significant body of opinion across Europe that is much less sanguine about sharing data at this scale than NHS England is at the moment.
Perhaps the only thing one can say with any certainty is that public concern about the security of its data is likely to increase over the next few years. Unless care.data takes these doubts seriously we may all live to regret its understandable ambitions.
Paul Hodgkin is founder and chair of Patient Opinion, a website on which patients, service users, carers and staff can share their stories of care across the UK. Patient Opinion is a not-for-profit social enterprise based in Sheffield.
Until 2011, Paul also worked as a GP and has published widely including in the BMJ, British Journal of General Practice and the Guardian and the Independent. Follow him on Twitter @paulhodgkin.