Shaping ideas is, of course, easier said than done. Bombarding people with ads only works to a degree. No one likes being told what to think. We grow resistant to methods of persuasion that we see through—just think of ads of yesteryear, and of how corny they feel. They worked in their day, but we’re alert to them now. Besides, blanket coverage isn’t easy to achieve in today’s fragmented media landscape. How many channels can one company advertise on? And we now fast-forward through television commercials, anyway. Even if it were possible to catch us through mass media, messages that work for one person often fail to convince others.

Big-data surveillance is dangerous exactly because it provides solutions to these problems. Individually tailored, subtle messages are less likely to produce a cynical reaction. Especially so if the data collection that makes these messages possible is unseen. That’s why it’s not only the NSA that goes to great lengths to keep its surveillance hidden. Most Internet firms also try to monitor us surreptitiously. Their user agreements, which we all must “sign” before using their services, are full of small-font legalese. We roll our eyes and hand over our rights with a click. Likewise, political campaigns do not let citizens know what data they have on them, nor how they use that data. Commercial databases sometimes allow you to access your own records. But they make it difficult, and since you don’t have much right to control what they do with your data, it’s often pointless.

This is why the state-of-the-art method for shaping ideas is not to coerce overtly but to seduce covertly, from a foundation of knowledge. These methods don’t produce a crude ad—they create an environment that nudges you imperceptibly. Last year, an article in Adweek noted that women feel less attractive on Mondays, and that this might be the best time to advertise make-up to them. “Women also listed feeling lonely, fat and depressed as sources of beauty vulnerability,” the article added. So why stop with Mondays? Big data analytics can identify exactly which women feel lonely or fat or depressed. Why not focus on them? And why stop at using known “beauty vulnerabilities”? It’s only a short jump from identifying vulnerabilities to figuring out how to create them. The actual selling of the make-up may be the tip of the iceberg.