From The Economist;

Researchers who have listened to teenagers talk frankly report that, for many, porn is the main source of sex education. Even those who have not viewed it have heard plenty about it from friends. It is shaping their expectations of sex—and what they go on to do.

For those who grew up in a pre-internet world, our main source of sex education was probably some kind of combination of what we heard from older kids, big brothers and sisters (whether ours or our friends), and maybe a few magazines; a world of naivety and personal experiences, but very much a world where "in the land of the blind, the one-eyed man is king".

If porn is moving from print to online, then there is a significant change in the impact it has on sex education. What you are presented with is no longer an editorially selected image, but an algorithmically selected set of videos. Instead of a handful of titles on the shelves of a newsagent — all aimed at the same audience — there is an infinite selection of content, on permanent rotation.

And what is determining the selection? An algorithm, which takes on board what everyone is clicking on. Assuming that the Pareto Principle holds (and there isn't really any obvious reason to think it wouldn't), then the 20% of heaviest users are accounting for 80% of the traffic — so its their choices and preferences that are mainly determining what everyone else is exposed to.

Including those teenagers for whom porn is the main source of sex education.

Maybe those 20% of most active users are actually the teenagers who are finding their way around the world, and this isn't really an issue — just a feedback loop. But I would suspect that the real valuable audience that these kinds of algorithms are being optimised for would be a different kind of audience — at the very least, old enough to have a credit card of their own. And probably not looking for the kind of videos that I would want to be the backbone of my childrens' sex education.