This article suggests that machine learning and collaborative filtering are incapable of producing healthy recommendations. I beg to differ, the New York Times may not like the result but they work for the vast majority of users on any service with too much content to manually curate.
Society has had laws in place to prevent children from viewing things they should not be (inappropriate movies, magazines, etc).
On the internet it is much more difficult, of course, and we can't realistically expect some shady offshore site from implementing age checks, let alone recommendation algorithms. But Google is a public, respected company from a first world country that claims to be promoting social good (which, of course, is marketing BS, and even if it weren't I would not want their idea of social good, but still). You'd think that they would invest some effort into not showing inappropriate content to kids at least. But no, they throw up their hands and go on ideological witch hunts instead.