Big Data talks see good turnout

Warwick students received interesting perspectives from leaders in the field of data collection and surveillance, as One World Week organisers and the Politics, Philosophy and Economics (PPE) Society brought in specialists for talks last week.

The Breach of Privacy in the Digital Age, the first of a series of talks organised by One World Week Forum, saw a good turnout.

All of the speakers were open in their interaction with the students including David Gioe, despite still retaining his commission as a Naval Reserve Intelligence Officer.

He pointed out that surveillance was necessary in order to prevent security breaches. When the topic of terrorism was brought up, he added that people didn’t see how effective surveillance could be in foiling terrorist plots because these details are hardly ever made public.

Laura Brandimarte, a post-doc fellow at Carnegie Mellon University, spoke about her research in the behavior of privacy.

She discussed statistics such as the fact that when people were asked the question if they had ever posted drunk pictures of themselves on Facebook, a large percentage of those that said yes also said that they wouldn’t hire someone who posted such pictures.

Nick Pickles, director of Big Brother Watch, offered his view from the political perspective, stating that we as the public have to make decisions regarding how much we want to reveal about ourselves at every step of the way.

Jean-Michel Corrieu, director at IBM International Business Solutions, was an interesting addition to the group as he offered the corporate opinion on the topic and how data collection, while intended to be anonymous, can do damage.

He contrasted two incidents. One was an incident in which a supermarket brand predicted that a teenage girl was pregnant based on her consumption pattern and sent her diaper coupons, causing embarrassment and controversy.

The other included a different brand, which allowed for free shipping of products if women submitted a pregnancy certificate and was very successful. He also discussed IBM’s in-built security measures at every point of data collection.

All of the speakers agreed that at the decision ultimately came down to the public making a trade-off between their privacy and need for security.

A follow-up to the talk was Big Data, organised by the PPE Society.

The talk featured Viktor Mayer-Schönberger, professor at Oxford Internet Institute and co-author of ‘Big Data: A Revolution That Will Transform How We Live, Work, and Think’.

Coining the term “more-messy-correlations” to describe Big Data and its usage, he discussed the implications of using huge amounts of data as compared to smaller data inputs.

A researcher at University of Toronto hospital, by analyzing vast amounts of medical data, predicted that premature babies have a higher chance of getting a fatal infection if their vital signs stabilise, as opposed to the original idea that fluctuating signs would signal this.

He also stressed the importance of correlation against causality while analysing data and the value that data holds, suggesting that its value is not exhausted once used; it may be reused for further analysis.

One example cited by him was of Rolls Royce predicting the likelihood of damage to engine parts by collecting data from the engines and doing predictive maintenance, thus reducing the chances of machine failure in-flight.

Datafication was another aspect he addressed. For instance, IBM has come up with a datafied floor, which tracks movements across it. Applications include prediction of consumer behavior in commercial institutions and assistance of elderly people.

However, he added: “By endangering free will, we run the risk of eliminating responsibility.”

He suggested problems that such analysis could lead to in the future, one of them being predictive punishment. Punishment without proof, but based on the analysis of behavior patterns, would be questioning human will and freedom.

He also pointed out that the only control to prevent misuse of so much information is data user responsibility.