Big data has been big news in recent months, not least with the scandal following Cambridge Analytica's abuse of Facebook data to persuade people to vote a certain way in the US elections.
But more generally, people have become aware of the incursion of big data in their lives, as they see targeted adverts on their social media feeds and the use of algorithms in their selection of films and books online.
Such is the emergence and use of data in our everyday lives that some interested parties are getting together to discuss the use of big data and whether there should be ethical guidelines in the interaction between big data and its commercialisation.
Digital society
In Switzerland, a collection of concerned parties, including Munich Re and Swiss Re, are getting together under the Digital Society Initiative to build a series of ethical guidelines over the use of big data among the insurance companies of the world.
Dr Markus Christen, managing director of the Digital Society Initiative at the University of Zurich, said: "We have companies that at least give the impression they realise [managing data] is not only a purely technical and legal point of view – they've started to take ethical considerations into account. They are saying, 'We might be able to do X but do we want to do X?'"
He added: "People themselves in the company, at least to some extent, don't want to do ethically bad stuff. They're starting to realise what's possible, but [are thinking] maybe they shouldn't do anything that's possible."
Key points
There are two areas that use of big data in financial services – marketing and risk assessment. Marketing is where companies use the data obtained about customer behaviour and know-how to advertise certain products, even at certain times of the day.
Risk assessment is deciding who is a reliable customer and offering them better deals on products or services, or conversely avoiding the high-risk clients altogether.
The insurance industry has been using its data for decades in order to decide who is high risk and who is low risk, and using algorithms to determine how to set premiums.
Michael James, head of technical architecture of Altus, said: "In the general insurance industry they have huge amounts of data, but the problem they have is do they have the right permissions.
"In general insurance they tend not to use it for customer acquisitions because they don't have the permissions. But they can use it to generalise it, so that personality type A is high risk and personality type B is low risk." The insurance sector can also use postcodes to determine whether one is a safe driver, for example.
The fear among some in the financial services sector is what happens if, or when, Amazon ventures into financial services. Looking again at the area of insurance, the concern is that Amazon has so much specific data about its customers that it could produce much more targeted premiums, based on what it knows, that no one else knows.
Mr James said: "Alexa is collecting data on households every day. If it knows I'm not someone who claims ever and runs a really safe house it could probably offer me a really good deal. If Amazon could match my levels of risk, I would be quite happy.
"The only downside is that it means people who are high risk are going to have problems [getting cover]."
Not everyone is in agreement with him. Dr Christen said: "Markets in insurance and banking are highly regulated and companies have to fulfil a lot of legal requirements to operate in that market. It might be more likely that [the likes of Amazon] will sell the data to banks and insurance companies."
Information harvest
One company has been harvesting unique data from its clients, and using it, to a degree, to offer customers financial products. This is Sainsbury's, which owns Nectar, and is developing its financial services arm Sainsbury's Bank.
Mark Hunter, chief data officer of Sainsbury's Bank, said: "We're building a database to add value for our customers and come up with innovative products and a proposition that means customers get more out of the group.
"It's the type of car, or how you drive on the road; these are the indicators of risk or price that we offer back to the customers. With home insurance, it's the type of property, and it's the number of people in the house – this is all the same information that insurers have access to.
"We can add a little bit of data that gives us an edge that we share back to our customers by offering a discount."
He added that Sainsbury's can use the Nectar data in the same way that banks use data from their customers to offer discounts on financial products, or discounts on one's shopping if one buys a mortgage, for example.
Clearly other companies such as Netflix and Amazon use algorithms based on previous behaviour to sell new products or services, but the challenge for any company using big data is contextual integrity.
Context matters
Dr Christen said that we may feel happy about giving away our data in a health-related context, but if that data is then used in an alternative context, it could be compromised. He said: "If we are with friends we [behave] in a certain way; if we are with employers we [behave] in another way."
The problem is that these different contexts are in danger of getting mixed up in the way that data is passed around.
For example, in Switzerland if one is planning to rent an apartment, one might be asked for criminal records that show you have not paid rent in previous places.
Dr Christen said: "To a certain degree it's ethically acceptable that we have to provide this information, but you could have the fear that certain people will never get certain offers any more."
Data has been collected and used in financial services for decades, the difference now is that there is much more available, and it is more tightly controlled under General Data Protection Regulation.
The question is whether in the future it can still be handled sensibly, and whether companies that collect data have learnt from the Cambridge Analytica saga.
Melanie Tringham is deputy features editor at Financial Adviser and FTAdviser.com