You are here

Dr Brian MacNamee speaks to The Irish Times about monetising personal data

Insight's Dr Brian MacNamee spoke to Marie Boran for an Irish Times article about the perils of monetising your personal data. In the piece Macnamee says, “One of the big things in all of this is the [power of] combination: what companies are trying to do all the time is to combine these different pieces of data and connect it to something else that was given somewhere else and start to build up this much richer picture of you,” says Dr Brian Mac Namee, a researcher in machine learning and statistics at the Insight Centre for Data Analytics at University College Dublin.

Mac Namee gives the example of recent research that connected Mosaic data on consumer classification (ie education, income level, media consumption) with location streams from people’s phones, adding a new dimension of data to the consumer: where they spend most of their time.

“Someone has attached another data source to an existing one and now they have this much richer picture of you. With individual little bits of data that you think ‘That’s not so bad’ and sometimes we get a bit surprised and realise those different bits of data that seemed innocuous by themselves, you’re not 100 per cent happy with the picture somebody gets to have of you when they put those pieces together.”

There is a disconnect, says Mac Namee, between the data we hand over up front and how that data is handled in the background: “It flows out to that coffee shop example: that the front face of who you are notionally giving your data to – a business or website you trust – trust that’s not really who you’re giving the data to because there’s a line of hidden third parties standing behind them.”

But surely these connections all add up to provide the personalisation we have become accustomed to. Most of us enjoy product recommendations, personalised searches and notifications of cinema releases tailored to our tastes. Isn’t profiling the price we pay, whether we get paid for our data or not?

“DuckDuckGo argue against that, “ says Mac Namee. “They say this idea of profiling and building those detailed profiles to use for things like advertising – their argument is they can make a very good search engine without doing any of that and at the same time they can make their money doing this by serving ads that are related to whatever you happen to be looking for at the time. They get 80 per cent of the personalisation they need from just looking at things you happen to be typing in at the time.”

Mac Namee’s analogy to put all of this in perspective is a visit to the local restaurant. You visit once or twice a week and the waiter gets to know you. Next time you come in, he says: ‘It’s great to see you. We have this lovely merlot that we think you’ll enjoy.’ This is fantastic and you’ll certainly be back.

“The flip side of that is if you discovered that the waiter had strolled across the road to the bank and said: ‘You know, next time Brian is in here I wouldn’t loan him any money, he’s a total lush; he’s in here two or three times a week putting away two bottle of merlot each time.’ You obviously wouldn’t like that.”

This behind-the-scenes data-sharing is the reason why Facebook, Google and other data giants can infer undisclosed information such as marital status, sexual and political orientation, education level, and even your IQ level from your fast-food choices (Facebook found a positive correlation between lovers of curly fries and a high IQ).

When it comes to your digital assets, for now, whether you’re selling it or giving it away, caution is advised because you don’t know how long the line of third-party brokers is and how many dots are being connected behind the scenes.

You can read the full article here.