A few years ago, Kord Davis found himself in meeting after meeting with technologists and product marketers who were -- without exactly knowing it -- discussing the ethics of big data. The meetings went something like this: The technologists would introduce new stuff they could do with data; one product marketer would call it cool and another would call it creepy.
"I realized, in the absence of a common vocabulary and framework for talking about ethical questions in the business context, we revert back to our own personal moral codes," said Davis, former Cap Gemini consultant and author of Ethics of Big Data: Balancing Risk and Innovation, during an O'Reilly Media webinar.
Ethics aren't an easy conversation to have in a business setting, Davis said. The topic is huge, messy and, yes, personal, and it tends to get tacked on to a pile of pressing obligations. In other words, it's not a topic that's easily dispatched. But in the absence of legislation that keeps pace with technological advancement, Davis believes it's a conversation that should happen -- especially among those practicing big data analytics.
"There are a lot of innovation values that can be generated from big data. What we want to do is minimize the risk," he said. The definition of risk will differ from company to company, of course, but Davis believes there are steps every business can take to develop a code of ethics for how they handle data: Be explicit about company values vis-à-vis data; be willing to discuss how to build policies around data privacy, customer identification and data ownership; align business actions with those values -- and prepare to have disagreements.
In the long run, Davis said, the benefits of transparency on how a company collects and uses data will outweigh the number of disagreements and uncomfortable conversations. "There are organizations out there that are clear about their values and communicate them openly and explicitly to their customers," he said, "and they have huge brand loyalty." Just ask Patagonia, Ben & Jerry's and Newman's Own.
Another benefit: A common set of values can actually increase the pace of innovation. "It turns questions from 'Should we do this?' to 'How can we do this?'" Davis said.
Personal data and the Fortune 50
And by the way, don't look to big companies for the answer on the ethics of big data. As part of Davis' research, he read the data-handling policies for the Fortune 50 and found the language to be inconsistent and revealing because of what they didn't say.
Not a single policy explicitly stated the company was selling personal data. Instead, more than half -- 34 of the 50 -- stated they wouldn't sell personal data without consent. Also, not a single policy explicitly stated the company would refrain from buying personal data. Instead, 11 of the Fortune 50 disclosed the practice of purchasing third-party data. But are the companies checking to make sure the data they've bought has been disseminated with the data owner's consent?
"It raises the question: If it's not OK to sell something, how is it OK to buy it?" Davis pointed out.
Transparent and explicit
It's easy to say businesses need to be more transparent and explicit when it comes to data collection and usage, but following one's own rules and building trust with customers is difficult, Davis acknowledged. And while he didn't have all of the answers, he suggested looking at Umpqua Bank, a regional bank in the Pacific Northwest, for insight.
The Data Mill
Relational databases are far from dead -- just ask Facebook
Ten case studies on big data in a nutshell
MetLife fires up JSON and Synapse to recruit rock-star developers
"They have a consumer-friendly brand, they play pop music and give out cookies," he said. "And they're very open."
Customers are sent a one-page data-handling policy that's written in what Davis called human language. It describes what the bank does with personal information and how it's kept.
On the flip side, according to Davis, is Orange County in southern California and its longstanding legal battle with the Sierra Club over the public's access rights to its geographic information systems (GIS) maps. The county refused to make the GIS maps available to the environmental organization, claiming its GIS databases were computer software and as such did not fall under the Public Records Act. Orange County ultimately lost the battle and may now have to pay $1 million in legal fees.
Get social, CIOs
There's no way around it. For all you CIOs who have not taken to Twitter and other social media outlets, it's past time. Why? First, CIOs who don't use social media on a personal level are less likely to understand what the enterprise is doing with it and why it's important. Second? The business already believes you are knee-deep in social media. This is according to Elden Nelson, analyst with the Stamford, Conn.-based consultancy Gartner Inc., who bases his thinking on a recent survey he helped conduct.
"In almost all social media activities, with [the category of] 'collect data' being the only exception, the business side of the enterprise was more likely to believe IT was playing a role than the IT [side of the enterprise]," Nelson said during a webinar where he shared the results. IT social media activities cited by the survey's business respondents included maintaining tools, defining and implementing a strategy and analyzing the data.
Since business already acknowledges IT has a role to play in its social media programs, Nelson suggests CIOs make them a priority. "IT can move off the sidelines and start helping organizations move from the fairly ad hoc way we're seeing social media programs planned, deployed, used and maintained to a system that can be used and integrated within the enterprise as a whole."