One of the important topics schools around the world are teaching on Big Data is ethics. According to Patrick O'Sullivan, Professor of Ethics at Grenoble Ecole de Management, "Big Data is here to stay simply because computing and the internet are here to stay. So it becomes a question of drawing lines around what is acceptable".

O'Sullivan sees two sets of ethical issues. The first addresses the limits of privacy. Does anybody have the right to collect data on others without their permission? The second issue, which O'Sullivan says is related to the first, is whether people are aware of what they're doing when they do give their consent.

"In some cases, they are not," says O'Sullivan. "In other cases, they should be aware, but they rarely are, because they are being asked to sign a set of conditions nobody ever reads through unless they can stand three hours of sheer boredom reading through a ridiculously long and complicated legal document."

Nick Pickles, director at Big Brother Watch in the UK, explains the problem: "Consumers are handing over information in the belief that it will be used for one purpose, but on the very odd page 14, in small print, lies a statement that gives the company rights to share the information with a third party. So the data is acquired under a false pretext."

Pickles points out that in signing up for a loyalty card, people might agree for one use of the data. Most people are happy to have the supermarket use information on their shopping patterns to generate discount offers. "But what if the supermarket draws inferences from the information and sells that to an insurance company?" he asks. "Few consumers would be happy to know that the store they shop at communicates to insurance companies that they should pay high health insurance premiums because they buy too many doughnuts."

Pickles says that sometimes people agree to have data collected on them in the belief that the information is not linked to their identity. "The trouble is, most of the time, data is useless without linking it to a person. In many cases when companies say consumers remain anonymous, what's really true is that the consumer remains semi-anonymous, or semi-identifiable."

According to Pickles, part of the problem is re-identification, which is the process of combining data sets to identify people who were previously anonymous, or pseudonymous. Pickles says: "Re-identification is an unprecedented opportunity of Big Data and the damage to privacy and confidentiality is potentially huge."

Patrick O'Sullivan believes that in some ways Big Data can be thought of as any other product: "The seller has to ask whether the product will be use for moral purposes. In the case of Big Data, any company passing on consumer data to another company should be asking whether passing the data on will result in a further invasion of privacy."

[Part 1 - What every organisation needs to know about Big Data]

[Part 3 - What UK organisations need to do now with Big Data]