This came up in a TED talk about the ethics of online advertising a while back, but I can't recall who the speaker was. They talked about the transformation of advertising in the age of big data, and how marketing efforts have shifted from benign, widely-encompassing efforts in print and tv to highly-targeted, tailored initiatives to drive sales. The implication being that by using consumer data, we could be building systems that are capable of targeting those who struggle with impulsive behavior or are easily co-opted (think old people and scam callers) and not even know it.

It really comes down to how you view things like addiction and impulsive behavior, and how those things impact free will. The fact that we don't really know with any kind of certainty that people who struggle with these things are protected when using the internet raises a lot of questions. Should people with these struggles just not use the internet?