Obviously, if you build a killer robot, it's totally okay as long as you have no idea how it works.
The thing that's opaque is this function that the program produces. You can really only know how the function works experimentally.
If you buy a handgun, and sometimes that handgun goes off without the trigger being pulled due to faulty design, then that's the one time the handgun manufacturer is actually liable.
It's not as if you get to tell the algorithm "Hey, I want Facebook to be a positive experience" or opt in to content.
When you have an interest like religion or politics and the algorithm starts showing you anti-vaxx material, that isn't the user in control. It is more like that faulty firearm.
For a handgun both factors are widely understood in the majority of the adult population.
If I build a robot that shoots a man when I press a button - and I press the button - then I am still to blame.
If I build a robot that runs a bunch of algorithmic code and it shoots a man - I am still to blame.
So the orginator of the algorithm (Facebook) is to blame, in full and without reprieve. The algorithm did not exercise free will.
Had Facebook made it so that the algorithm was transparent, or at least easily disabled then they might have some defense. They did not do this.