I'd mock the terrible quality of the comic, but that's like mocking the prototype orphan grinding machine for only being able to mangle limbs.
I have to say that the argument over whether AI is/isn't Copyright infringement is not the part that worries me the most. No, what has
me concerned is stuff like this:
"Oh I don't see the issue here," I hear many say. "Of
course the soulless Capitalist companies will flock to the development of tools that maximize profit and productivity, even at the cost of human suffering/happiness/classism/etc."
No, that's not the point. The point here is
not that AI models are being developed to identify and purge the lowest performers. The point is that the ones training those AI models, the ones giving these AI models their "morals," so to speak,
are those same soulless corporations. Questions about Turing tests and eventual debates about what constitutes "sentience" aside, when companies swap their completed, purpose-trained AI models between themselves, the
Weights that are inherent in those models are going to be ones trained to isolate that company's preferred flavor of "deadwood," so to speak, and that means that any kind of decision-making these models (or their derivatives) perform in the future will have that bias bred right into them.
I guess what I'm saying here is something
I'm positive I've already discussed--that it's fine to let the AI do identification, but
not okay to let an AI make actual
decisions. THAT part should be left to actual people.
--Patrick