Using AI to convert a story to audio format?

[Otto26]

I have a number of stories which I would like to convert to audiobook format. I'm looking at text to speech utilities and services and they are ALL using AI. I'm willing to accept the second rate output I would get. But I'm not willing to invest in a service and then find out I'm violating Lit policy. I've reviewed the AI policy documents I can find (images and generated text) and believe I am on the right side of those. I've read accounts from other Lit authors saying they've done this in the past.

Is anyone aware of any Lit policies I would be violating by using a text to speech tool to convert stories I've written?

 

[MillieDynamite]

Yeah, no AI is the site-wide policy. While I don't think it says audio, AI art, and writing are forbidden, and I bettcha, so is AI voice.

 

[Otto26]

The AI prohibition is aimed at generative AI which creates bodies of text or image. This is about using a text to speech tool. That's substantially different.

 

[AwkwardMD]

The AI prohibition is aimed at generative AI which creates bodies of text or image. This is about using a text to speech tool. That's substantially different.

You would be generating content at a pace not possible under normal conditions without significant monetary resources. It is the same.

EDIT: the only difference is that the AI is replacing a voice actor rather than the writer. These are both important creative jobs whose value should not be underestimated.

 

[AwkwardMD]

For the record, I use AI text to speech to hear my stories out loud. I think it's a great tool for me, to process my work through a different sense. It's now a permanent step in my editing process.

 

[MillieDynamite]

I do the same thing, and so does my editor. Hearing how something sounds is important and as critical as the other edit. Grammar mistakes (unintentional) are glaring.

For the record, I use AI text to speech to hear my stories out loud. I think it's a great tool for me, to process my work through a different sense. It's now a permanent step in my editing process.

 

[Otto26]

You would be generating content at a pace not possible under normal conditions without significant monetary resources. It is the same.

EDIT: the only difference is that the AI is replacing a voice actor rather than the writer. These are both important creative jobs whose value should not be underestimated.

Ah. So what pace am I allowed to work at?

 

[baffling8929]

Ah. So what pace am I allowed to work at?

Seven

 

[AwkwardMD]

Ah. So what pace am I allowed to work at?

I don't know how much sympathy you think you will garner by purposefully misunderstanding others in order to present a gotcha moment, but I can assure you it's less than you might hope.

 

[Otto26]

I don't know how much sympathy you think you will garner by purposefully misunderstanding others in order to present a gotcha moment, but I can assure you it's less than you might hope.

Not my intent. But here, I'll try a different tack.

I think you're misunderstanding the intent of AI bans. It's more about copyright issues than anything else. There are multiple lawsuits against AI companies alleging copyright infringement and some of those cases are likely to end up in the Supreme Court and shape future legislation as well as exposing publishers to lawsuits or massive (costly) clean-up efforts. The contention is that the work of humans has been stolen and used to train AI. The works are derivative rather than original. Which is why the site policy on visual art also states that you can't take someone else's content and re-purpose it for your use and claim ownership.

I think this because I've read all the Lit AI policy I can get my hands on and what I have read leads me to believe that using text to speech does not violate any site policies. But I might be wrong. It happens more frequently than I'd like. So I asked about site policy here where others can offer an informed opinion. But what you gave me seemed to be an uniformed, and incorrect, opinion. It's not about how fast someone produces content.

 

[MillieDynamite]

@Otto26 If you want to know if you can or can't, send Laurel a private message and ask her if AI voice is allowed for audio stories. Dah, it isn't rocket science or even grade school math hard. It clearly states that they only want content produced by human beings. That seems to cover, to me, any form of AI here, what so ever.

 

[Otto26]

If you want to know if you can or can't, send Laurel a private message and ask her if AI voice is allowed for audio stories. Dah, it isn't rocket science or even grade school math hard.

I bother the site admins as an absolute last resort. If the answer to my question is already published I'd like to find it before I bother them. And if I don't find anything that says it's not allowed then I'll put one of my published short stories into audio format and submit it, with a clear explanation of exactly how it was produced. They'll kick it back or they won't. They kicked back the image I tried to publish, testing the limits of what is allowed, and all I said was 'Thank you'.

 

[Voboy]

@Otto26 If you want to know if you can or can't, send Laurel a private message and ask her if AI voice is allowed for audio stories.

This is the way.

I don't like to "bother" the site admins either, OP, but I doubt this would "bother" them any more than having to reject your submission and then deal with it twice (once when you submit, again when you resubmit with a Note To Admin), accompanied by the same PM exchanges you could more productively make before you submit.

I'm with @AwkwardMD and @MillieDynamite in thinking that any and all forms of AI are verboten, but I also know I have no idea what the answer to your question is. Laurel does have that answer. Asking her won't hurt anything. Might even help, by providing a definitive answer from which others could benefit.

 

[ElectricBlue]

I think using a speech generator as part of an edit process is one thing, using it to publicly to present something in place of a real human is another.

Mind you, my word clone "voice" is obviously a robot, and sucks at being anything useful for a decent reading. I can read out loud if I need to.

I wouldn't use a more sophisticated version, like the one Scarlett Johansson got taken down. That's just wrong, for all the AI reasons it's wrong.

As the others say, it's a question for Laurel. The FAQ on AI are high level and global, and it seems to me that looking for an audio loophole is trying to dodge the policy.

 

[Katie_Mae]

It has been allowed in the past, but that's not to say it would be allowed now.

- As an example, Laurel allowed an entry to be published in early 2023 which had a significant portion (380 words out of 750 words) of it generated by generative AI (the author informed her that it was AI generated), but I think that would no longer be allowed.

This thread https://forum.literotica.com/threads/audio-conversion-through-ai.1577656/ discusses AI audio, including a link to a story here on lit with audio generated by AI. Again, the story was posted a couple of years ago, so that isn't to say it's no longer allowed.

AI text to speech programs may require payment per so many words, or limited words per month for free plans, so depending on which program you plan on using it might be time consuming or costly to generate a whole story (though I'm sure there's also free ones out there, depends what you're looking to use). Unless you're also looking at posting the audio elsewhere, it might be worth while asking first. [Edit, to clarify there's no current policy that I know of specifically banning AI text to speech, but that doesn't mean Laurel would approve it now)

 

[MillieDynamite]

Strange game, the only winning movies seems to be not to play. How about a nice game of chess?

 

[AwkwardMD]

 

[cadeauxxx]

[spam removed]

You can build your own voices on there and change pronounctuations. I use it as a way of testing dialogue now to see how my words sound spoken aloud by a voice.

It's also a great tool for musicians with stripping out layers of instruments from songs to create backing tracks, if you're a musician.

 

[Otto26]

Never heard back from the admins on this subject so I used a generic, free text-to-speech tool to convert one of my stories to audio and submitted it. I included a note to the admin which stated I was using a text-to-speech tool to convert one of my existing, published stories to audio. It was accepted for publication. So, hooray.

I've requested the story be pulled. Now that I know this is an acceptable tool I'm willing to invest in improved TTS capabilities. I'll re-submit the story with higher quality audio. I'll also correct some other mistakes. The most important of which will be to give my submission a title which makes it clear it's an audiobook (Audiobook - Story Title). That will help listeners who want a first person narrative filter out my stories.

I'm excited about this. It won't work for all my stories, of course. I'll still need to find voice actors for some of my works.

 

[AwkwardMD]

Never heard back from the admins on this subject so I used a generic, free text-to-speech tool to convert one of my stories to audio and submitted it. I included a note to the admin which stated I was using a text-to-speech tool to convert one of my existing, published stories to audio. It was accepted for publication. So, hooray.

I've requested the story be pulled. Now that I know this is an acceptable tool I'm willing to invest in improved TTS capabilities. I'll re-submit the story with higher quality audio. I'll also correct some other mistakes. The most important of which will be to give my submission a title which makes it clear it's an audiobook (Audiobook - Story Title). That will help listeners who want a first person narrative filter out my stories.

I'm excited about this. It won't work for all my stories, of course. I'll still need to find voice actors for some of my works.

Just be ware. Literotica is a literary site first, and that's what gets the most attention from the site's sole moderator. Pictures, audio recordings, and art are supported, but are not prioritized. There is a strong chance that your story getting approved for publishing had less to do with tacit approval of your method and more to do with apathy and disinterest. Slipping through the cracks, if you will. If your submission gets reported and later rejected for AI, despite your best efforts to be transparent, I wouldn't be surprised. That's a non-zero chance.

That being said, it certainly looks like I was wrong.

 

[Otto26]

Just be ware. Literotica is a literary site first, and that's what gets the most attention from the site's sole moderator. Pictures, audio recordings, and art are supported, but are not prioritized. There is a strong chance that your story getting approved for publishing had less to do with tacit approval of your method and more to do with apathy and disinterest. Slipping through the cracks, if you will. If your submission gets reported and later rejected for AI, despite your best efforts to be transparent, I wouldn't be surprised. That's a non-zero chance.

That being said, it certainly looks like I was wrong.

Understood. I don't assume any of the risk or liability the site operators do so my opinions matter very little. The quality of the reading is slightly better than monotone, but not by much. If you want to hear how bad it is you can listen before it's removed (which I have already requested). It's gone. I don't think this kind of audiobook reading represents a threat to human creators.