We can look at the intention of the developers? https://news.ycombinator.com/item?id=24579498
The law already concerns itself with determining intent (hence murder vs. manslaughter). I don’t see why similar criteria and methods cannot be applied here. We can assess the objectively bad psychological consequences of social media and look for evidence that establishes moral certainty about the intent.
I am not impressed with manufactured skepticism.
Let the user decide. The evil algorithms require the "algorithms" such as "all my subscriptions, completely, in chronological order" to be removed. If "all ragebait" or "maximize profits for facebook" are just other buttons, I have no objection, but I do not think they will be popular.