With those examples though, how would we know ahead of time that they "shouldn't be explored?" They sure looked interesting and maybe even potentially beneficial a couple decades ago.

Now, of course, we know those algorithms warp regular users (and by extension societies). Or... maybe they don't? Some research has suggested that just putting this many people in direct communication with each other is the root cause of the problems we see. There could be other ways to fix those without shutting down the internet. How would we know without more exploration?