no subject

[firecat]

The Nanowrimo peeps shouldn’t have referenced ableism, classism, and privilege in their recent post about AI.

But I don’t understand why a person would be categorically opposed to all uses of whatever-is-being-called-AI-this month in a creative pursuit.

(The majority of my use of AI has been getting suggestions for cat names, so that tells you how much you should pay for this opinion.)

There are ways to use AI, for example as a prompt or name generator, that are not “getting it to write your whole novel.” Why would someone object to such uses?

The Nanowrimo posts everyone is piling on already say that using AI to do the actual writing misses the point.
It’s a problem for creators that so far no copyright law covers what AIs can consume, but that’s a separate issue from whether they have legitimate uses.

I seem to be at variance with most of my opinion bubble about this issue. Feel free to tell me what I’m missing.

Context: https://www.404media.co/email/3d9698b2-8c2b-41e7-bea4-7a1ac6916159/

Comments

[violsva]

I think there's a kind of hair-trigger effect. After a writer has been hearing about all of the terrible training of AI, and the terrible output, and flooding submissions, and that it's already putting writers and artists out of work and lowering their wages, and also it turns out it's destroying the environment, and so on and so forth, for nearly two years now, no one is giving any reference to AI the benefit of the doubt. (And not just writers!)

Given that AI is a completely meaningless marketing term, maybe they should. But it's reasonable that people are not willing to take the time to carefully separate out what AI they think might be okay and what isn't.

Also, a lot of people don't think the question of legitimate uses is actually separate from the issue of training datasets. Many people are not willing to use any output from a system they know was trained on unlicensed copyrighted works, even if they might be okay with the use itself.

[firecat]

Thanks for your views and the links!

Also, a lot of people don't think the question of legitimate uses is actually separate from the issue of training datasets. Many people are not willing to use any output from a system they know was trained on unlicensed copyrighted works, even if they might be okay with the use itself.

Yeah. I think it’s fine to eschew tools marketed as AI for ethical reasons, and I want to see conversations about the ways it is causing harm. Or if people think Nanowrimo is making the statements they are in order to pander to their corporate supporters, I want to know that too.

But I see something else going on in a lot of the conversations about this.

[violsva]

I generally feel people should take advantage of the fact that the internet makes it easy to cite your sources!

[firecat]

That is an excellent point. Me too.