FCC to restore #NetNeutrality and also, hey, to protect our online activity from scrapping--be it for #ArtificialIntelligence training without our consent, or for even more nefarious purposes.
When the public comment period opens, for your own interest, comment--the vultures will, by the thousands, using bots/AI, so we need to show up and make sure our voices are heard, and our interests protected.
I know it's super tempting, especially when you barely make any money off your art, but please realize that in the long run, you'd be eliminating your own role: if #AI "can narrate it" there's no reason why it couldn't "write it".
And yes, this also goes for covers; you can't draw the line at screwing over other creators.
@herhandsmyhands@romancelandia It’s a tricky one. I have people with dyslexia or limited sight ask about audiobooks all the time. I feel badly explaining why I don’t have audio.
I experimented with a few AI narrations when GooglePlay came out with it, but stopped. It didn’t quite feel right to me.
But paying for my catalog to be narrated live? There goes my income, spiraling away. Despite audio growth, it’s still too tiny of a % to be worthwhile for most of us.
That's addressed by Bree in the third screenshot; she and Donna have an extensive backlist they can't afford to put on audio, because they can't afford to pay a human.
It's not tricky. It's a personal choice--like using AI covers because they "can't afford human-made art".
Some people will rationalize it, others won't, but the ethics and longterm consequences are actually clear.
@herhandsmyhands@romancelandia Thank you for sharing this, the narration was something I hadn’t thought of. I’m currently raising money to pay an artist who will design a book cover for me. Is it as cheap as AI? No, but I knew from the get go that I wouldn’t choose that route. I’d rather wait and do it right.
@herhandsmyhands@romancelandia It's funny - I don't have a problem with basic TTS. I feel like that's a decision on the user's end based on their own personal accessibility needs, made with the understanding that the result will be imperfect.
But once the end product is being marketed as an "audiobook" (as opposed to an ebook the reader may run through a screen reader if they wish)...
Those are two very different products and I'm not happy about efforts to blur the line.
I love well-narrated audiobooks (and I have very definite ideas about what constitutes good narration), but text-to-speech provides basic accessibility without forcing a whole class of creative people off the field.
@herhandsmyhands@romancelandia I think the reason they are so scared about having to pay is, that they don't really expect to actually make money from AI, they wish to make it appealing enough that some poor smuck pays through the nose for it. And having to pay for the material would both cut the promised near infinite payout they promise, and might become too expensive for them to be able to maintain the con for long enough.
@herhandsmyhands@JiSe@romancelandia They fought real hard to put copyright above first sale and private property rights. they literally passed a law to eliminate 1st sale doctrine: the DMCA.
now that the situation they made is hampering them from committing theft (in their own words!) they suddenly have a change of heart. allegedly.
@herhandsmyhands@romancelandia anyone who uses LLMs as any part of their writing or publishing process instantly loses all respect from me. That's it. They're dead to me.
I'm likely missing something, but why not generate an audiobook via local software? Maybe I don't understand what "AI narration" is? Is it AI trained on other people's voices in the usual plagiaristic way?
I guess I have no problem with mechanical narration for authors who can't afford human narration and can't do it themselves, but maybe I'm missing something? I get that mechanical is 2nd best, but are there other issues?
@tjradcliffe
Text-to-speech is what you mean by local software; TTS is an assistive technology, not "second best" but different from a voice actor narrating a book.
AI audio (or "virtual voice") uses voice actors' stolen voices (including books they've narrated) to 'train' a simulacrum, in the same way that ChatGPT is trained through stolen intellectual property.
Add comment