Public Domain

Comments about LibriVox? Suggestions to improve things? News?
Post Reply
daisycampion
Posts: 16
Joined: November 10th, 2021, 1:47 pm

Post by daisycampion »

I just wanted to share that I think it's unbelievable, when you take a step back, to think that every day more books are going into the public domain. Every single day. That's something to wake up and be joyful about. :clap:
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

True indeed. It is also brilliant that there are a lot of people aware of free licenses like Creative commons and GPL and so on. Freeware. Linux is brilliant, it is something you can take the time to invest in without fear that one day someone will stop producing that software and you can't use it, OR YOUR KNOWLEDGE about it, anymore.

I knew a lot of tricks and tips for windows 95 and a few for XP, and absolutely nothing for whatever is the current version of windows. No point. Learning windows is all pointless. Obsolete. Investing in a stock that will inevitably go to zero.

Long live freely shared works.

As a post script, I regularly get out old software for linux, I download it and use it even though it's way way out of date, because I can. It works and I can use it to draw 3d or do sound, or whatever. It's fabulous and it won't ever happen for commercial software. They force you to use ever more invasive garbage spyware that cannot do what it pretends and claims to be able to do. Garbage.

Best things in life are free.
lightcrystal
Posts: 1202
Joined: October 22nd, 2021, 10:55 pm
Location: Melbourne with kangaroos

Post by lightcrystal »

GettingTooOld wrote: June 1st, 2023, 4:12 am True indeed. It is also brilliant that there are a lot of people aware of free licenses like Creative commons and GPL and so on. Freeware. Linux is brilliant, it is something you can take the time to invest in without fear that one day someone will stop producing that software and you can't use it, OR YOUR KNOWLEDGE about it, anymore.

I knew a lot of tricks and tips for windows 95 and a few for XP, and absolutely nothing for whatever is the current version of windows. No point. Learning windows is all pointless. Obsolete. Investing in a stock that will inevitably go to zero.

Long live freely shared works.

As a post script, I regularly get out old software for linux, I download it and use it even though it's way way out of date, because I can. It works and I can use it to draw 3d or do sound, or whatever. It's fabulous and it won't ever happen for commercial software. They force you to use ever more invasive garbage spyware that cannot do what it pretends and claims to be able to do. Garbage.

Best things in life are free.
Yes. The public domain gives incredible reading and listening resources. There are some treasures on the internet archive.

Apple is tightening its leash on customers. Logic Pro and Final Cut are both subscription products on the iPad. Surely that's a dry run for doing the same to the desktop versions.
I'm a geek. I wreck my computer. Then I fix it.
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

lightcrystal wrote: June 1st, 2023, 5:57 am Apple is tightening its leash on customers. Logic Pro and Final Cut are both subscription products on the iPad. Surely that's a dry run for doing the same to the desktop versions.
Often the case. Even if a licence is free, but the software is copyright, it is the same problem, they can and will eventually stop issuing licences. Companies go broke or issue new products after building a mousetrap version of cheese with a new features updated version for a low low price.

Everything that has a beginning has an end, so too with licenced software. Sooner or later all the effort you put into learning it is made useless.

Free software, which is a gift to all from the writer is something you can keep on using. For decades. I'm sure some of the software for drawing in 3d or image manipulation in 2d which I use is probably over 10 years old. Perhaps some of the software I use for programming today is over 20. I can still use it and no-one can stop me. 8-P

I don't know what logic pro is, but final cut sounds like openshot, which I used today.
annise
LibriVox Admin Team
Posts: 38542
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

daisycampion wrote: May 14th, 2023, 12:28 am I just wanted to share that I think it's unbelievable, when you take a step back, to think that every day more books are going into the public domain. Every single day. That's something to wake up and be joyful about. :clap:
Actually, PD is like racehorses who all have the same birthday so a 1 year horse could in theory be anything from 1 day to 365 days old. The date probably varies in different countries but the concept is the same - it's August here I think but possibly 6 months earlier in the Northern hemisphere
So books only become PD on January first of each year.

Anne
GettingTooOld
Posts: 416
Joined: October 19th, 2021, 3:28 am

Post by GettingTooOld »

annise wrote: June 1st, 2023, 7:09 am [..]PD is like racehorses[..]
It can be very fustrating if the one you have your eye on is too slow to come around, you wish it would HURRY UP it's taking forever !!!

Whereas a good specimen will get crowds of people excited, especially if it has a great ending !
Rel22
Posts: 129
Joined: September 1st, 2022, 9:58 pm
Contact:

Post by Rel22 »

RE: viewtopic.php?t=98823

I'm so glad that the Librivox admin team has taken a firm stand in not allowing the use of A.I. to takeover the works of human volunteers/voices (literally) on this platform. However, as private interests and large corporations utilise AI to disproportionately take advantage of anything available on the internet (and presumably public domain) for private profits and control over future intellectual property, I'm wondering now about the lack of protection (ultimately) that is given to narrators recording audio on these public domain books and releasing them back into the public domain. Likewise, AI can scrap for all the human-written summaries and catalogs after the fact, couldn't they? And those AI companies would unapologetically take them to build up their databases for not only financial but indiscriminate social/cultural gains. Wouldn't the very same parties who would have taken over at the beginning (if no bars existed) now capture enormous benefit from this back end too? For the voiceovers, they may scrap all the human voices and sell them or use them to train AI, or whatever they wish, because of 'public domain'--I'm supposing. I haven't really been following so closely (though now I probably should) on the tech space on AI, but there is a feeling of a kind of fatalistic end to our natural rights in exploitation in just a matter of time unless we have some additional guard rails in place.

I was a little (perhaps naively) surprised hearing for the first time of a case where a voiceover artist's recording was lifted by a big company and used by AI to mimic and sell her voice without her permission (though apparently she may have not realised the contract she was signing had a clause that did say to this effect that ultimately the payment given was in exchange for waiving her rights for any future use including AI). But now what about in the case of public domain? (Arguably even the spirit of 'public domain' is broken by AI, though policy makers never had to be faced with something like artificial intelligence before now, I again suppose.)

What is the wisdom or the legal ideas on this topic?
It makes me hesitate (as a narrator) to continue recording for the public domain and not because I'm of any person of significance but because I'm wondering if I'm contributing to something increasingly vulnerable and AI is trying to take advantage of all of us and it hasn't been fully regulated yet. Are there any other policies that should be considered or ideas/comments on this subject?

I'm open to any and all views...
:thumbs:
Rel
annise
LibriVox Admin Team
Posts: 38542
Joined: April 3rd, 2008, 3:55 am
Location: Melbourne,Australia

Post by annise »

It has been going on for a long time - you don't need AI to do it. Audible has many books read by LV readers - they are just a platform for the people who add them. It's like all charity work - do you refuse to support starving people because others steal the food and sell it ? You have to decide. Many people benefit from our teamwork, do we stop because others legally try to make money from it?

Anne
redrun
LibriVox Admin Team
Posts: 2756
Joined: August 11th, 2022, 8:32 pm
Contact:

Post by redrun »

Rel22 wrote: August 24th, 2023, 10:41 pm I was a little (perhaps naively) surprised hearing for the first time of a case where a voiceover artist's recording was lifted by a big company and used by AI to mimic and sell her voice without her permission (though apparently she may have not realised the contract she was signing had a clause that did say to this effect that ultimately the payment given was in exchange for waiving her rights for any future use including AI). But now what about in the case of public domain? (Arguably even the spirit of 'public domain' is broken by AI, though policy makers never had to be faced with something like artificial intelligence before now, I again suppose.)
Another angle you might consider, Rel. First, a recap of a couple points I think you alluded to:
The current crop of "AI" systems are not created in a vacuum - as you note, they are products of the big companies, and they exist to make those companies money. But they also need to be "fed" a lot of training data to be useful at all, and while it's still being argued, copyright law could prevent them from training using whatever data they can get their hands on.

But... copyright isn't going to stop them:
The companies that currently make these "AI" tools have the advantage of working at a very large scale, with lots of money to buy whatever they want. And now they know they want it. If they have to pay, run, or even buy a popular web site, so that they can use user posts as training data, they will. The terms of service for these sites either do or will say that users give the site operators permission to copy whatever they post, and those sites will sell that permission on to these large companies. Zoom recently (and surprisingly) said they won't sell or use anything for this particular purpose, but most sites will instead do what Reddit is doing: adding a paywall to make sure they get their slice of the pie.
(Edited to add: sites that are run for money will likely get bought or paid to change their terms of use. LV never will be - there's nothing to buy, and our books are already not covered by copyright.)

I'm afraid the genie is out of the bottle, and even if these companies have to wipe what they have so far and re-train using only content they pay to access, they will all do it. The possible abuses of these tools - deepfakes, (easier) plagiarism, and (easier) misinformation - are here to stay. There are people out there who will train new models, illegally or no, that will enable this abuse. But the positive uses - automated translation, summarization, pattern-finding and audio accessibility, among others - are also here to stay, and they're not reserved for just big companies. Or at least, they shouldn't be.


Respectfully disagreeing with one point on "violating the spirit" of public domain:
On the other hand, the public domain exists so that any use, whether previously imagined or not, can be made of works after their creators have been remunerated. Our culture is full of remixes and derivative works based on the stories that our ancestors told, and our technology is all based on inventions that were or would have been patented in their day, but that we can now make free use of. Without this system, no new story or invention containing any piece of existing tales or technologies could be made... without first finding the heirs of the original authors or inventors of each piece you want to use, and getting all of their permission to do something new with it.

How the public domain might make for better "AI":
Having more works in the public domain means there's more training data available... but not just available to these few big companies. The current large companies will have these "AI" tools from now on, and even if they keep a lid on possible abuses, there will also be smaller "enterprises" that ignore the laws and make abuses easier. But having more (legally usable) training data freely available means that as the technology advances, smaller and smaller companies, groups, or even individuals can make their own "AI" tools. Even if most people use the tools from the large companies, just knowing that there's an alternative can be a powerful thing. If they abuse their market power, they might just find their customers going to a more ethical company... but only if that company has a good product in the first place.

I considered the down-sides of putting my own recordings in the public domain, but I think the human uses now, and potentially beneficial "AI" uses in the future make it worth the risk in my case. But I certainly wouldn't judge anyone for whom it's different. You'll notice I've also chosen not to put my real name on my account, for exactly the reason of making it harder to abuse my voice (as tied to me as a person) in some future scheme that I can't guess at. Privacy and identity are tricky enough in a world of public access, but I think they would be at least as bad in a world of backroom-negotiated copyright deals.
Rel22
Posts: 129
Joined: September 1st, 2022, 9:58 pm
Contact:

Post by Rel22 »

annise wrote: August 24th, 2023, 10:58 pm It has been going on for a long time - you don't need AI to do it. Audible has many books read by LV readers - they are just a platform for the people who add them. It's like all charity work - do you refuse to support starving people because others steal the food and sell it ? You have to decide. Many people benefit from our teamwork, do we stop because others legally try to make money from it?

Anne
The speed and disproportionate benefit that AI wields right now is what makes it feel like are we inadvertently contributing in a naive way? Just to be clear though, I love what Librivox.org stands for, it's what drew me into it in the first place. I was relieved to know such a platform existed that really values the human touch on things, and humans reading to other humans. That's why I wanted to contribute. But the people who may be benefiting from the teamwork here, I'm just wondering aloud if it is a losing battle ultimately unless we, for example, worked simultaneously on something more extreme like had AI also built that guards us. Without that confidence of knowing we have that safety net, it doesn't mean that we quit but the pause gives me reason to think there has to be more discussions and maybe more extreme ways in which we give a platform like this greater protection than just to keep going and 'hope for the best'.

Thank you so much for your view and opinion, Anne. I really appreciate the discussion.
:thumbs:
Rel
Rel22
Posts: 129
Joined: September 1st, 2022, 9:58 pm
Contact:

Post by Rel22 »

redrun wrote: August 25th, 2023, 7:45 am
Rel22 wrote: August 24th, 2023, 10:41 pm I was a little (perhaps naively) surprised hearing for the first time of a case where a voiceover artist's recording was lifted by a big company and used by AI to mimic and sell her voice without her permission (though apparently she may have not realised the contract she was signing had a clause that did say to this effect that ultimately the payment given was in exchange for waiving her rights for any future use including AI). But now what about in the case of public domain? (Arguably even the spirit of 'public domain' is broken by AI, though policy makers never had to be faced with something like artificial intelligence before now, I again suppose.)
Another angle you might consider, Rel. First, a recap of a couple points I think you alluded to:
The current crop of "AI" systems are not created in a vacuum - as you note, they are products of the big companies, and they exist to make those companies money. But they also need to be "fed" a lot of training data to be useful at all, and while it's still being argued, copyright law could prevent them from training using whatever data they can get their hands on.

But... copyright isn't going to stop them:
The companies that currently make these "AI" tools have the advantage of working at a very large scale, with lots of money to buy whatever they want. And now they know they want it. If they have to pay, run, or even buy a popular web site, so that they can use user posts as training data, they will. The terms of service for these sites either do or will say that users give the site operators permission to copy whatever they post, and those sites will sell that permission on to these large companies. Zoom recently (and surprisingly) said they won't sell or use anything for this particular purpose, but most sites will instead do what Reddit is doing: adding a paywall to make sure they get their slice of the pie.

I'm afraid the genie is out of the bottle, and even if these companies have to wipe what they have so far and re-train using only content they pay to access, they will all do it. The possible abuses of these tools - deepfakes, (easier) plagiarism, and (easier) misinformation - are here to stay. There are people out there who will train new models, illegally or no, that will enable this abuse. But the positive uses - automated translation, summarization, pattern-finding and audio accessibility, among others - are also here to stay, and they're not reserved for just big companies. Or at least, they shouldn't be.


Respectfully disagreeing with one point on "violating the spirit" of public domain:
On the other hand, the public domain exists so that any use, whether previously imagined or not, can be made of works after their creators have been remunerated. Our culture is full of remixes and derivative works based on the stories that our ancestors told, and our technology is all based on inventions that were or would have been patented in their day, but that we can now make free use of. Without this system, no new story or invention containing any piece of existing tales or technologies could be made... without first finding the heirs of the original authors or inventors of each piece you want to use, and getting all of their permission to do something new with it.

How the public domain might make for better "AI":
Having more works in the public domain means there's more training data available... but not just available to these few big companies. The current large companies will have these "AI" tools from now on, and even if they keep a lid on possible abuses, there will also be smaller "enterprises" that ignore the laws and make abuses easier. But having more (legally usable) training data freely available means that as the technology advances, smaller and smaller companies, groups, or even individuals can make their own "AI" tools. Even if most people use the tools from the large companies, just knowing that there's an alternative can be a powerful thing. If they abuse their market power, they might just find their customers going to a more ethical company... but only if that company has a good product in the first place.

I considered the down-sides of putting my own recordings in the public domain, but I think the human uses now, and potentially beneficial "AI" uses in the future make it worth the risk in my case. But I certainly wouldn't judge anyone for whom it's different. You'll notice I've also chosen not to put my real name on my account, for exactly the reason of making it harder to abuse my voice (as tied to me as a person) in some future scheme that I can't guess at. Privacy and identity are tricky enough in a world of public access, but I think they would be at least as bad in a world of backroom-negotiated copyright deals.
Thank you redrun for your considerable response and different angle. I much appreciate it.

I have to agree with your point that 'the genie is out of the bottle' indeed. That is exactly part of the problem. But I question if a beautiful, well-intentioned platform and team of human volunteers (not just here but I'm sure countless other similar ones), adhering to the 'spirit' of human endeavor, is enough with the introduction of AI.

The 'spirit of public domain' I refer to is the spirit intrinsic to if we were all using human endeavors to use and reuse stuff and eventually reciprocate by donating to the public domain and the cycle repeats. But AI is kinda alien, isn't it? because it's goes beyond human and functions at a level that we aren't privy to nor can we understand as we would of one another, plus the benefits of the remix/reuse of it all goes into the hands of a few? The copyright law only works if humans can prove it in a court of law and at this rate with the introduction of AI stealing even copyright material, we would now need AI to sift through AI's vast ability to merge copyright stuff with non-copyright stuff to even determine if copyright has been violated. I have read of a computer program developed by a whiz kid in university doing something along those lines in order to give professors a leg up on knowing if something produced in class has been done with AI or not. That certainly gives us some hope in other applications than education but it's definitely not enough. AI isn't being properly or fully regulated (which is wishful that it would be completely regulated because of the big companies' interests involved). So 'until then', we do have to protect ourselves individually but also consider how would we also protect such a great platform?

Indeed one way of protecting yourself individually is what you alluded to yourself in giving yourself a pseudonym. But the very fact that you even have to do that (though of course countless others do it too and not just on this platform) implies that it feels like we don't have any real firepower collectively other than these small protections we do for ourselves, while Rome burns. I just question if we're all inadvertently feeding the very monster which we have been hoping to ward off at getgo without also discussing other brighter ideas from the greater community that values these platforms to provide not just perspective but something more concrete we could all do to preserve the nature of such a platform and its aftermath.

I think we all need to talk about it more and more to figure it out. But I'm glad to know you and others are willing to talk about this.
:thumbs:
Rel
ej400
Posts: 5132
Joined: September 24th, 2014, 10:26 am
Location: Minnesota
Contact:

Post by ej400 »

FWIW, I personally will continue to contribute to Librivox because:

1. It's fun
2. To have an entire site out there with the thousands of audiobooks, knowing they are read BY humans, is a big thing in my honest opinion.
3. I don't like the sound of robotics speaking, no matter how far they progress, or who takes voices and tries to steal the works of humans, I still will only listen to librivox audiobooks, because they are FREE and because I know the community that creates them.

Personally I don't see AI taking over for audiobooks, and if they do... Amazon audiobooks is a big thing because amazon is large cooperation. But the thing is, this is a FREE place for public domain works created by volunteers, and so I don't sense a competition.
Post Reply