Big Data, big danger?

ORIGINAL INTRODUCTION

In recent years, with advances in data-storage capacity and accessibility, more and more information about us is being captured and analysed by government agencies and private companies. While there is some nervousness about this, the proponents of ‘Big Data’ argue that the benefits outweigh the risks to privacy.

For example, the benefits of having huge pools of data from medical records to examine the pros and cons of medical treatments or the dangers of certain lifestyle choices – allowing hitherto unrecognised relationships to be identified and new hypotheses to be formed – are said to be worth the risk that medical confidentiality will be breached. We are told that if the police and security services have access to mobile phone, credit card and travel card data, then it will be easier to prevent crime and terrorism.

However, while computers can do the heavy lifting in collating and analysing data, the choices about which data to collect are made by humans. For example, why choose to collect data from Oyster cards instead of using passenger surveys? Another criticism is that researchers’ prior assumptions are given the veneer of objectivity and authority simply by being stated in statistical form. The way that data is used as a replacement for political principle is also troubling. It is far harder to hold to the principle that people should be free to drink alcohol as they please if opponents have stats and graphs that seem to confirm a strong link between alcohol and violent crime. As a result, politicians may endeavour to justify policies using the supposedly unimpeachable evidence of Big Data rather than on more philosophical grounds.

There are also concerns for individual liberty. Data can tell us much about trends and tendencies across populations. But there are fears that data may be used preemptively, to identify individuals who are more likely to commit crime, become obese, fail at school or suffer an early death – leading to premature or unnecessary state intervention.

Regardless of the sophistication of computer technology, ultimately decisions must be based on human judgements. For example, statistician Nate Silver achieved fame by correctly predicting the outcome of the 2012 US presidential election in 49 out of 50 states. But, as Silver himself noted, his predictions were based on applying mathematical techniques to a range of information, including the opinions of experts. Expert judgement is still required to enable Big Data to fulfill its potential.

Should we question the hype around Big Data? Is it really true that the benefits of Big Data outweigh the dangers? Should we place limits on data collection to protect individual liberties? Even in areas where data is put to benign use, could over-reliance on algorithms impede the process of human judgement? Is it time to recognise the limitations of Big Data and put the stats in their place?

SPEAKERS
Sheila Bird
programme leader, MRC Biostatistics Unit, Cambridge Institute of Public Health

Yann Bonduelle
partner, UK Consulting leader, Data & Analytics team, PwC

Marion Oswald
head, Centre for Information Rights, University of Winchester

Sandy Starr
communications officer, Progress Educational Trust; webmaster, BioNews

CHAIR
Timandra Harkness
journalist, writer & broadcaster; presenter, Futureproofing and other BBC Radio 4 programmes; author, Big Data: does size matter?