LA is fundamentally ethically different from the analysis done by online vendors because of power differentials. The relationship between you and your English 101 instructor, or your institution of higher education as a whole, is much different than your relationship between you and Google or Facebook or Amazon.
When online, we often operate under the assumption that our activities are occurring in our bedrooms, so to speak. We don’t always consciously operate as if our online activities are actually occurring out in the scrutinizable open – although we should be aware of that, if we read the fine print in the terms of agreement. Yet, if they are aware, most people are OK with knowing that Google/Amazon/Facebook is mining their every tweet. Most of us don't know anybody working for those corporations, and even if we did, we know that we are one of multimillions...our online actions are just a drop in the big data bucket.
However, the dynamics between a student and the institution of higher education in which he/she is enrolled is quite different. In this case, the people analysing their data trail may actually KNOW them, and may have the power to award grades, offer or withhold a job reference, or dole out scholarships, work-study jobs, or internship leads. Now, our data is subject to the institutional gaze, and those eyes do have the power to reward and punish. It is a power differential tilted markedly toward the institution. Learning analytics may help a student persist along the path towards college completion, but I believe that the analysands' feelings about being scrutinized by individuals in positions of power (including their instructors) at a university would be much different than their feelings about Google or Facebook mining even that same pot of data.