Toward an ethics of generative AI
re: the ethical bit
A lot of anti-AI people liken GenAI to guns and think the tech is similarly ethically fraught down to its foundations. Guns, with few exceptions, are made to put holes in living things despite non-killing and self-defense uses.
For guns, it's hard but not impossible to make a case the non-killing capabilities outweigh that function. For LLMs, I think it comes down to whether you believe people who've found all kinds of good uses for them. If so, there's ample value outside whatever one thinks of the companies building it.
And that question drives the debate: are the people finding value in LLMs dupes who don't realize they're idiots, or are they actually useful if operated sensibly with a mind toward their current limitations? If the former, LLMs are evil; if the latter, humans can have a little GenAI, as a treat.