I read this article in The NYT: Mapping, and Sharing, the Consumer Genome, Natasha Singer. It is about data giants creating a typical consumer. They know who you are, where you live, what you do, your age, race, sex, weight, education level, politics, buying habits, ...

Technically it might be challenging to extract consumer behavior models out of the massive information, but it has a Big Brother aspect that cannot be neglected.  

As a machine learning system provider, I think of the potential of misinterpretation and even misuse. There are so many exciting industrial data mining and machine learning applications that's errors or excessive use do not carry cost for humans and we will not knowingly provide mlf to data miners working with human profiles.