But, you see, it is not me who is offering the heroin, it is just this machine which guesses the user's preferences...
From what I see Facebook already implements automated content detection but you know it’s not perfect, content slips through and it’s often in the gray area.
Would you like your account to be blocked because you joked about something? Bots don’t have sense of humor.
Do you want some ibuprofen? Sorry, that’s a drug. Request dropped.
(Heck I’m even afraid this comment gets marked as spam because of the keywords)
Of course I wouldn't like my account to be blocked because I posted a slightly edgy joke. But would we care if an edgy memes page gets marked as unsuitable for recommendations? Probably no. Would it be a problem if Facebook blocked an Ibuprofen ad to some users that are deemed as not suitable? I wouldn't care, in all honesty.
If limiting the extremism slippery slope of targeted reccomendations requires to mark extremist content as unsuitable for reccomendation, then I personally wouldn't mind. Facebook already does it to some extent, but its algorithms are either not good enough, or driven predominantly by other factors