All Things Techie With Huge, Unstructured, Intuitive Leaps

Data Mining And Ethics - Revenue Over Rights?

I was reading a WIRED magazine article where the author said that his liberal arts degree on his resume was viewed in the same way as a face tattoo in a job interview for an investment banking position. However, a liberal arts degree is probably the qualification of an up and coming job title for data mining companies -- the CEthO or the Chief Ethics Officer. Once data mining and machine learning gets to the next level, coupled with the Internet of Everything, there will be huge privacy and ethics issues to contend with.

The big question of ethics that will arise, is "Is is okay to make money off information about people that I glean from mining my data?"

Several far-fetched but not so far-fetched scenarios come to mind. I am reminded of the data mining done by Target Stores when they deduced that a 15 year old girl was pregnant by her purchases of face cream combined with a certain brand of vitamins.  Suppose that I was a data miner for a drug store chain, and I could find a strong correlation between a person buying certain antacids and a few months later being diagnosed with an ulcer requiring expensive stomach surgery. A health insurance company would be highly interested in knowing that. Should we sell the information on people that we discover? It would be an incredibly lucrative revenue stream.

Ethics was never a question in the good old days of business. McDonald's grew their fast food empire by putting toys into their Happy Meals and creating an obese America by targeting and hooking the children. One nutritionist noted that Chicken McNuggets in a Happy Meal were nutritionally worse than deep-fried cake. At the time, it was seen as a slick marketing move.

Data mining is in the same stage that McDonalds was fifty years ago. It is a solution looking for a problem to monetize. So it is like the Wild West until legislators, sober second-thought minds and ethicists added some groundwork rules to the field of endeavor. I personally know of a data miner who set up in a Caribbean jurisdiction that has little to no privacy laws. Big companies ship him their data complete with personal information, and he ships back everything bit of intelligence that he finds. His revenues are out of this world. It is almost the ethical equivalent of having consumer products made in the Third World by child labor. I can see the days, where data becomes a commodity that is bought, sold and traded. There will be export laws on data -- especially data with personal identification information in it.

But there are ethical questions closer to home. Suppose that my employer expects me to mine data, and I discover an untapped revenue stream that is extremely easy to exploit. Do I tell my employer, or give my notice and create a start-up to exploit that situation? What is the ethical course of action?

We have a long way to go with with applying ethics to data mining. Ethics is a lot like beauty -- it is in the eye of the beholder. I am reminded of a story that a shopkeeper told regarding ethics. He said "I was closing the store and a customer came in at the last minute and made a large purchase with a hundred dollar bill. After he had left and I closed the shop, as I was counting the money, I noticed that there was actually another hundred dollar bill stuck to the bill that the last customer had tendered. Immediately a question of ethics arose. I was wondering if I had to tell my business partner or not."

And that perfectly explains why we need some ethical boundaries in data mining.

No comments:

Post a Comment