Big Data Potential Can Be Fulfilled With Distributed Governance – Sandy Pentland

Sandy Pentland‘s closing comments in a dialogue at the recent Overseas Development Institute in London reminded me of Sir Thomas More‘s advice to King Henry VIII of England, “too much power in the hands of one person will surely lead to tyranny.” As if to prove the point, Henry had Thomas More executed for saying it.

Pentland’s stewardship of the MIT Human Dynamics Lab puts him at the forefront of thinking around the enablement of society through big data and was invited to the Overseas Development Institute in London highlighting the potential of big data in a world where we often know very little of the trends that lead to human suffering, even though we have the technology.

In reference to a UN report on the big data Data Revolution which was commissioned by Secretary General Ban Ki-moon, Pentland highlights the things that the report pointed out, leaving the trailing thought of distributed governance to reduce the risk of power abuse

Official statistics and big data are perhaps a marginal utility, in the near future in London. They are (however) a critical utility in most places in the world, there are millions of children that die every year, because we don’t know.

There are millions of people that die from ethnic violence, from all sorts of causes, from infectious disease because we don’t know and it’s too expensive and too slow to use traditional techniques to be able to establish those statistics.

So, when you want to put things off, you need to ask, it’s gonna take this amount of time and you can just multiply by a few million then you get the number of children that you allow to die because you’ve delayed the process and I would like you to think of it that way, there is an urgent need.

What is this sort of data we talk about when we talk about big data? The data we are most familiar with, we don’t think of as big data, for instance satellites, those that tell us about global warming or tell us about failure of food crops, about deforestation, GPS, there are maps everywhere now. That wasn’t true twenty years ago. Those maps help you get supplies to things, deal with infectious disease and so on, they are big data, they really are and they are critical to operating our world.

Add to those a number of other things, for instance, the books, the financial books of government, that’s pretty big government and I’d like to know where government spends our  money, wouldn’t you? Or the money of companies, not just the details but you know who gets what would be good to know.

Aggregate data would be a great place to start there that wouldn’t really bother anybody too much. Also mobility, cell phones, cars, things like that know where they are so you can actually look at flows of people and again, aggregates are not disturbing at all. We all use navigation aids in our cars and on our phones. That is aggregate big data. I’m sorry, you’re all users and you don’t feel bad about it because it’s this aggregate that doesn’t relate to any person.

Then there is also communication data, what communities talk to each other, what communities talk to other communities and with these you can do an amazing variety of things, you can look at economic equity, you can look at social isolation and equity, you can look at the progress of infectious disease in real-time which would have been useful in the ebola case in Africa, except the laws did not permit anybody to use that data, so perhaps thousands of people died, right?

You can do things like estimate likely hood of ethnic violence by looking at physical community overlap, it’s been shown that you can do a pretty good job of that, wouldn’t that be something that’s nice to address. You can even do mental health. Not in detail and of particular people but you can estimate the frequency of metal health problems, which are incidentally the worlds biggest health problem in terms of lost days of active life, it’s not cancer, it’s not violence, it’s not other things, it’s mental health.

So, we have the ability to address these huge needs, that can’t be addressed by traditional data. What we need to operationalise that is we need public and private cooperation, I don’t know if they are partnerships but most of this data is collected by private organisations and so somehow that has to be converted into the public good and that’s a question of liability and cost and so on.

It also raises the problem of privacy and control and I think that the thing that we need to do, and this is work that comes from the world economic forum, is we have to give citizens much more control over data that is about them, so when someone is recording something about you, you not only ought to know about it, you have a say into what happens to it and what you get from it, so we have that challenge as well.

To be able to actually make systems like this work, we have another real challenge, which is that, to a certain degree, these are pipe dreams. We can demonstrate the in scientific terms, we can do the odd million people and show that it works but what we need to do is we need to experiment, we need to find living labs where we can try these things out and prove that they work and prove they don’t violate people’s privacy and if they do, change them.

We don’t have to sit here and think up the perfect solution in isolation of reality, what we need to do is take the pressing need that people are dying every day, try some of the less controversial things out, see how they work and if they work well, implement them on a larger scale and so on, so forth, the way we do with every new technology.

The one thing that’s special about this new technology though, that I want to highlight, is that there is a special danger that is rarely talked about and that is concentration of power. When you have an enormous amount of data you have the ability to be able to persuade people and corral people and force people and that’s not true of many of the other technologies and it’s particularly true of government because not only do they have a lot of data, they also have guns, that’s an unfortunate confluence.

So we need to think about ways to do this that are distributed, we need to think about ways that the power is not in any small group of people’s hands, it’s in large groups of hands and that even in times when there is turmoil, you might have autocratic governments, you have at least some protection because the data is in so many hands.





About Gary Donovan

Machine Learning and Data Science blogger, hacker, consultant living in Melbourne, Australia. Passionate about the people and communities that drive forward the evolution of technology.
Show Buttons
Share On Facebook
Share On Twitter
Share On Linkedin
Share On Pinterest
Share On Stumbleupon
Contact us
Hide Buttons