It's no secret that the implications of big data extend far beyond organizational benefits into the societal and ethical realm, but market researcher Gartner predicts that the improper use of big data analytics will cause half of all business ethics violations by 2018.
Organizations could suffer loss of reputation, wasted resources, competitive weakness and even legal sanctions as a result, Gartner said.
The best-known example in this area may be the oft-told case of Target's pregnancy-prediction algorithm, which led the retail giant to deduce that a 15-year-old girl in Minnesota was expecting a baby. It wasn't until the girl began receiving coupons for baby-related items that her family caught on.
"Many of these violations happen unwittingly," Duncan said.
It's a learning curve similar to what's often required when new technologies are introduced.
"If we go back 100 or so years to the very first motor cars, there was a guy who used to walk in front of each car with a red flag to warn pedestrians," said Alan Duncan, a research director for analytics at Gartner. "People didn't know how to interact with the 'driverless carriage.'"
There may be a fine line between what's acceptable and not. While most consumers have become comfortable with the idea of putting personal information on social media sites such as Facebook and Twitter, they're not necessarily comfortable when companies "scrape" that data for a better understanding of them.
For example, if an insurance company starts sending fitness-related messages to consumers who have recently bought Fitbit devices, those people won't always welcome them.
"We see a huge amount of risk for companies to go blindly into using the data," Duncan said. "The data is amoral -- there is no good or bad, it just is. It's the human choices we then make that make the difference, and that's what data ethics is."
Much of it is still a gray area, and that's largely because the legal rules haven't yet caught up.
Ethics are "the substantive principles that inform our laws," said Neil Richards, a professor of law at Washington University in St. Louis.
"The problem is, the law inevitably lags," Richards said. "We pass laws when we find a problem, but we don't usually pass laws to deal with problems that haven't happened yet."
Regarding big data, the law doesn't say much yet about what businesses can do, he said. At the same time, there are areas where people frequently say "'that's creepy,' or 'unfair' or 'sleazy.'"
It often comes down to ethical principles to help a brand establish trust and reputation and develop brand identity.
What bothered Richards about the Target case was not just that the company knew more about the pregnant girl than her family did, but also what it did in response to the backlash that inevitably followed.
"When they realized people were getting freaked out by getting formula coupons, they then started sending out a full page of coupons with formula prominently displayed but surrounded by 'guy stuff' like lawn mower blades," he said. "They were still using the data, but now there was a level of deception to it as well."
Gartner has a suite of best practices it recommends as a way to avoid these sorts of problems, and they figured prominently on the agenda for its Business Intelligence and Analytics Summit going on in Munich this week. The bottom line is that companies must be transparent and communicate with consumers about what they're doing with the data, Duncan said.
"If you're using social science tools like big data, you have to be a good social scientist," Richards agreed. "All the things academic social scientists and data scientists have been worrying about, like selection biases, companies now have to worry about, too."
And because the law still lags behind in this area, "you can't just hire a lawyer and say, 'give us compliance,'" he said. "The good news is, there seems to be a real competitive opportunity for the companies that can get out ahead on ethics."