Yes, I have called this The Bias Paradox.
If a non-biased AI could actually exist, it would be perceived as biased and we would attempt to fix the non-bias resulting in injecting bias back into the system.
However, the output is not pure as data input is not the only source for potential bias in the system. All of the AI systems are purposely biased to remove harmful content in some manner as perceived by those managing the system.
Which would require deciding if a biased AI could actually exist… which would require deciding that an AI of any sort could actually exist… which begs the most important question, have we decided that _any_ form of intelligence actually could (much less does in fact) exist, or is that notion just our confirmation bias short-circuiting our quite literally primitive brains into a suicide mode?
The most useful thing these artificially ignorant systems will do is help us learn how stupid we really are. I just hope that they don’t do so by resolving Fermi’s Paradox for us.
> I just hope that they don’t do so by resolving Fermi’s Paradox for us
Yes, that thought has occurred to me. I've also considered that the "Singularity" might be instead of the point AI gains sentience, it is more like the moment we lose ours :-) as we go forward on foolish endeavors we don't fully understand.
> it is more like the moment we lose ours as we go forward on foolish endeavors we don't fully understand.
Ahh, so the Singularity is that thing rapidly receding in the rear view mirror then?
This species started looking for intelligent life on other planets before first establishing to any reasonable degree of satisfaction that intelligent life exists on this one… foolish endeavors we don’t fully understand is our thing.
However, the output is not pure as data input is not the only source for potential bias in the system. All of the AI systems are purposely biased to remove harmful content in some manner as perceived by those managing the system.
https://dakara.substack.com/p/ai-the-bias-paradox