I'm a big fan of consent. I just think that in some circumstances it isn't realistic, and I'm afraid this is one of them. It's interesting to have service providers agree on a model of consent, but what that means is that they're going to have to agree on what algorithms are going to do. I actually don't think that's feasible. I'm a professor, so feasibility isn't usually a big deal with me. I usually put feasibility to the side and talk about principle, so here I am using my own argument against me.
First of all, I'm not sure that's feasible. Even if there was going to be agreement about what algorithms will do and what they won't do, most service providers don't want to disclose what their algorithm does because it's how they make their money. There's an intellectual property there that they don't want to share.
I worry that the more we build up people's idea that we shouldn't worry because we've taken care of consent, the more we'll lose sight of the fact that we're dealing with something that is so difficult to know. Maybe if we thought about regulating the processes by which information is being dealt with, then we could say we've created an environment where consent or informed consent actually makes some sense.