7 ms·

Adobe isn't trying to be ethical, they are trying to be more legally compliant, because they see that as a market opportunity. Otoh, artists complain about lega

Adobe isn't trying to be ethical, they are trying to be more legally compliant, because they see that as a market opportunity. Otoh, artists complain about legal compliance of AIs not because that is what they care about, but because they see that as their only possible redress against a phenomenon they find distasteful. A legal reality where you can only train AI on content you've licensed would be the worst for everybody bar massive companies, legacy artists included.

_bin_ 2y ago
Right, but "distaste" isn't grounds for trying to ban something. There are all kinds of things people and companies do which I dislike but for which there's no just basis for regulating. If Adobe properly licenses all their training data artists don't have a right to say "well i think this is bad for creativity and puts my job at risk, ban it!!!" Or more precisely, they have a right to say that, but no moral justification for trying to ban/regulate/sue over it. I hate Adobe's subscription model as much as the next guy and that's a good reason to get annoyed at them. Adobe building AI features is not.
- TeMPOraL 2y ago
  > Right, but "distaste" isn't grounds for trying to ban something. It isn't, but it doesn't stop people from trying and hoping for a miracle. That's pretty much all there is to the arguments of image models, as well as LLMs, being trained in violation of copyright - it's distaste and greed[0], with a slice of basic legalese on top to confuse people into believing the law says what it doesn't (at least yet) on top. > If Adobe properly licenses all their training data artists don't have a right to say "well i think this is bad for creativity and puts my job at risk, ban it!!!" Or more precisely, they have a right to say that, but no moral justification for trying to ban/regulate/sue over it. I'd say they have plenty of moral / ethical justification for trying to ban/regulate/sue over it, they just don't have much of a legal one at this point. But that's why they should be trying[1] - they have a legitimate argument that this is an unexpected, undeserved, unfair calamity for them, threatening to derail their lives, and lives of their dependents, across the entire sector - and therefore that laws should be changed to shield them, or compensate them for the loss. After all, that's what laws are for. (Let's not forget that the entire legal edifice around recognizing and protecting "intellectual property" is an entirely artificial construct that goes against the nature of information and knowledge, forcing information to behave like physical goods, so it's not unfair to the creators in an economy that's built around trading physical goods. IP laws were built on moral arguments, so it's only fair to change them on moral grounds too.) -- [0] - Greed is more visible in the LLM theatre of this conflict, because with textual content there's vastly more people who believe that they're entitled to compensation just because some comments they wrote on the Internet may have been part of the training dataset, and are appalled to see LLM providers get paid for the service while they are not. This Dog in the Manger mentality is distinct from that of people whose output was used in training a model that now directly competes with them for their job; the latter have legitimate ethical reasons to complain. [1] - Even though myself I am for treating training datasets to generative AI as exempt from copyright. I think it'll be better for society in general - but I recognize it's easy for me to say it, because I'm not the one being rugpulled out of a career path by GenAI, watching it going from 0 to being half of the way towards automating away visual arts, in just ~5 years.
  - skissane 2y ago
    > they have a legitimate argument that this is an unexpected, undeserved, unfair calamity for them, threatening to derail their lives, and lives of their dependents, across the entire sector - and therefore that laws should be changed to shield them, or compensate them for the loss. After all, that's what laws are for. Lots of people have had their lives disrupted by technological and economic changes before - entire careers which existed a century ago are now gone. Given society provided little or no compensation for prior such cases of disruption, what’s the argument for doing differently here?
    - TeMPOraL 2y ago
      Moral growth and learning from history?
      - skissane 2y ago
        There’s a big risk that you end up creating a scheme to compensate for technological disruption in one industry and then fail to do so in another, based on the political clout / mindshare / media attention each has - and then there are many people in even worse personal situations (through no fault of their own) who would also miss out. Wouldn’t a better alternative be to work on improving social safety nets for everybody, as opposed to providing a bespoke one for a single industry?
        TeMPOraL 2y ago
        > Wouldn’t a better alternative be to work on improving social safety nets for everybody, as opposed to providing a bespoke one for a single industry? Yes, but: 1) It's not really an exclusive choice; different people can pursue different angles, including all of them - one can both seek immediate support/compensation for the specific case they're the victim of and seek longer-term solution for everyone who'd face the same problem in the future. 2) A bespoke solution is much more likely to be achievable than a general one. 3) I don't believe it would be good for society for artists to succeed in curtailing generative AI! But, should they succeed, I imagine the consequences will encourage people to seek the more general solution that mitigates occupational damage of GenAI while preserving its availability, instead of having to deal with a series of bespoke stopgaps that also kills GenAI entirely. 4) Not that banning GenAI has any chance of succeeding - the most we'd get is it being unavailable in some countries, who'd then be at a disadvantage in competition with countries that embraced it. Again, I'm not in favor of banning GenAI - on the contrary, I'm in favor of giving a blanket exception from copyright laws for purposes of training generative models. However, I recognize the plight of artists and other people who are feeling the negative economic impact on their jobs right now (and hell, my own line of work - software development - is still one of the most at risk in the near to mid-term, too); I wish for a solution that will help them (and others about to be in this situation), but in the meantime, I don't begrudge them for trying to fight it - I think they have full right to. I only have problems with people who oppose AI because they feel that Big AI is depriving them of opportunity to seek rent from society for the value AI models are creating.
- skywhopper 2y ago
  In the context of encouraging art, it totally is! Copyright and patents are 100% artificial and invented legal concepts that are based solely on the distaste for others profiting off a creator’s ideas. The reason for them is to encourage creativity by allowing creators to profit off new ideas. So there’s no reason why “distaste” about AI abuse of human artists’ work shouldn’t be a valid reason to regulate or ban it. If society values the creation of new art and inventions, then it will create artificial barriers to encourage their creation.
  - bmacho 2y ago
    Yup, banning AI for the sake of artist would be exactly the same as the current copyright laws. (Also they are attacking AI not purely for fear of their jobs, but bc it is illegal already.)
  - _bin_ 2y ago
    Disagree. Authority is given Congress to establish an IP regime for the purpose of "promot[ing] the progress of science and useful arts". You would have to justify how banning gen AI is a. feasible at all, particularly with open-weight models; and b. how it "promotes the progress of useful arts." You would lose in court because it's very difficult to argue that keeping art as a skilled craftsman's trade is worse for its progress than lowering the barriers to individuals expressing what they see. I think bad AI makes bad output and so a few people are worried it will replace good human art with bad AI art. Realistically, the stuff it's replacing now is bad human art: stock photos and clipart stuff that weren't really creative expression to start with. As it improves, we'll be increasingly able to go do a targeted inpaint to create images that more closely match our creative vision. There's a path here that lowers the barriers for someone getting his ideas into a visual form and that's an unambiguous good, unless you're one of the "craftsmen" who invested time to learn the old way. It's almost exactly the same as AI development. As an experienced dev who knows the ins and outs really well I look at AI code and say, "wow, that's garbage." But people are using it to make unimportant webshit frontends, not do "serious work". Once it can do "serious work" that will decrease the number of jobs in the field but be good for software development as a whole.
- cratermoon 2y ago
  > Right, but "distaste" isn't grounds for trying to ban something I disagree. There are many laws on the books codifying social distastes. They keep your local vice squad busy.
  - _bin_ 2y ago
    I thought most people supported moving away from that and towards a more socially liberal model. If we're no longer doing that I have a whole stack of socially conservative policies I guess I'll go back to pushing. I don't think y'all really want to go down this road; it leads straight back to the nineties republicans holding senate hearings on what's acceptable content for a music album.
    - fc417fc802 2y ago
      Many laws come down to distaste at the root. There's usually an alternative angle about market efficiency or social stability or whatever if you want to frame it that way. The same applies in this case as well. For but a few examples consider laws regarding gambling, many aspects of zoning, or deceptive marketing. What's the purpose of the law if not providing stability? Why should social issues be exempted from that?
- computerthings 2y ago
  [dead]
- weregiraffe 2y ago
  >Right, but "distaste" isn't grounds for trying to ban something. https://en.wikipedia.org/wiki/United_States_obscenity_law https://en.wikipedia.org/wiki/United_States_obscenity_law
Riverheart 2y ago
“A legal reality where you can only train AI on content you've licensed would be the worst for everybody bar massive companies, legacy artists included.” Care to elaborate? Also, saying artists only concern themselves with the legality of art used in AI because of distaste when there are legal cases where their art has been appropriated seems like a bold position to take. It’s a practice founded on scooping everything up without care for origin or attribution and it’s not like it’s a transparent process. There are people that literally go out of their way to let artists know they’re training on their art and taunt them about it online. Is it unusual they would assume bad faith from those purporting to train their AI legally when participation up till now has either been involuntary or opt out? Rolling out AI features when your customers are artists is tone deaf at best and trolling at worst.
- Workaccount2 2y ago
  There is no "scooping up", the models aren't massive archives of copied art. People either don't understand how these models work or they purposely misrepresent it (or purposely refuse to understand it). Showing the model an picture doesn't create a copy of that picture in it's "brain". It moves a bunch of vectors around that captures an "essence" of what the image is. The next image shown from a totally different artist with a totally different style may well move around many of those same vectors again. But suffice to say, there is no copy of the picture anywhere inside of it. This also why these models hallucinate so much, they are not drawing from a bank of copies, they are working off of a fuzzy memory.
  - TeMPOraL 2y ago
    > People either don't understand how these models work or they purposely misrepresent it (or purposely refuse to understand it). Not only that, they also assume or pretend that this is obviously violating copyright, when in fact this is a) not clear, and b) pending determination by courts and legislators around the world. FWIW, I agree with your perspective on training, but I also accept that artists have legitimate moral grounds to complain and try to fight it - so I don't really like to argue about this with them; my pet peeve is on the LLM side of things, where the loudest arguments come from people who are envious and feel entitled, even though they have no personal stake in this.
spoaceman7777 2y ago
> Adobe isn't trying to be ethical, they are trying to be more legally compliant Is the implication of this statement that using AI for image editing and creation is inherently unethical? Is that really how people feel?
- mtndew4brkfst 2y ago
  For creation, yes, because of the provenance of the training data that got us here. It was acquired unethically in the overwhelming majority of cases. Using models derived from that training is laundering and anonymizing the existing creativity of other humans and then still staking the claim "I made this", like the stick figure comic. It's ghoulish.
  - skissane 2y ago
    There exist image generation models that were trained on purely licensed content, e.g. Getty’s. I don’t know about Adobe’s specifically-but if not, it seems like a problem Adobe could easily fix-either buy/license a stock image library for AI training (maybe they already have one), and use that to train their own model-or else license someone else’s model e.g. Getty’s
    - spookie 2y ago
      Well they do license the art they use, but in... let's say... "interesting" ways through their ToS.
    - bolognafairy 2y ago
      They are training using licensed images! That’s the thing! There’s some sort of ridiculous brainworm infecting certain online groups that has them believing that stealing content is inherent in using generative AI. I watch this all quite closely, and It’s chronically online, anime / fursona profile picture, artists. Exact same thing happened when that ‘open’ trust and safety platform was announced a few months ago, which used “AI” in its marketing material. This exact same group of people—not even remotely the target audience for this B2B T&S product—absolutely lost it on Bluesky. “We don’t want AI everywhere!” “You’re taking the humanity out of everything!” “This is so unethical!” When you tell them that machine learning has been used in content moderation for decades, they won’t have a bar of it. Nor when you explain that T&S AI isn’t generative and almost certainly isn’t using “stolen” data. I had countless people legitimately say that having humans have to sift through gore and CSAM is a Good Thing because it gives them jobs, which AI is taking away. It’s all the same sort of online presence. Anime profile picture, Ko-fi in bio, “minors dni”, talking about not getting “commissions” anymore. It genuinely feels like a psy-op / false flag operation or something.
no_wizard 2y ago
> A legal reality where you can only train AI on content you've licensed would be the worst for everybody bar massive companies, legacy artists included. Quite an assertion. Why exactly would this be true?
- drilbo 2y ago
  who else has would ever have a significantly large store of licensed material?
  - fc417fc802 2y ago
    Or alternatively, who else could afford the licensing costs?
dinkumthinkum 2y ago
I'm curious why you think it would be worse for everybody? This argument seems to depend on the assumption that if something makes AI less viable then the situation for human beings is worse overall. I don't think many actual people would accept that premise.
- crimony 2y ago
  It's worse only if AI turns out to be of high value. In that case only large companies that can afford to license training data will be dominant.
tbrownaw 2y ago
> Adobe isn't trying to be ethical, they are trying to be more legally compliant, Ethics (as opposed to morals) is about codified rules. The law is a set of codified rules. So are these really that different (beyond how the law is a hodge-podge and usually a minimum requirement rather than an ideal to reach for)?
zmmmmm 2y ago
The ship has sailed, but I can understand artists feeling that no matter how any AI is trained prospectively, it was only made possible because the methods to do so were learned through unethical means - we now know the exact model architectures, efficient training methods and types of training data needed so that companies like Adobe can recreate it with a fraction of the cost. We obviously can never unscramble that egg, which is sad because it probably means there will never be a way to make such people feel OK about AI.

_bin_ 2y ago

Right, but "distaste" isn't grounds for trying to ban something. There are all kinds of things people and companies do which I dislike but for which there's no just basis for regulating. If Adobe properly licenses all their training data artists don't have a right to say "well i think this is bad for creativity and puts my job at risk, ban it!!!" Or more precisely, they have a right to say that, but no moral justification for trying to ban/regulate/sue over it. I hate Adobe's subscription model as much as the next guy and that's a good reason to get annoyed at them. Adobe building AI features is not.

TeMPOraL 2y ago
> Right, but "distaste" isn't grounds for trying to ban something. It isn't, but it doesn't stop people from trying and hoping for a miracle. That's pretty much all there is to the arguments of image models, as well as LLMs, being trained in violation of copyright - it's distaste and greed[0], with a slice of basic legalese on top to confuse people into believing the law says what it doesn't (at least yet) on top. > If Adobe properly licenses all their training data artists don't have a right to say "well i think this is bad for creativity and puts my job at risk, ban it!!!" Or more precisely, they have a right to say that, but no moral justification for trying to ban/regulate/sue over it. I'd say they have plenty of moral / ethical justification for trying to ban/regulate/sue over it, they just don't have much of a legal one at this point. But that's why they should be trying[1] - they have a legitimate argument that this is an unexpected, undeserved, unfair calamity for them, threatening to derail their lives, and lives of their dependents, across the entire sector - and therefore that laws should be changed to shield them, or compensate them for the loss. After all, that's what laws are for. (Let's not forget that the entire legal edifice around recognizing and protecting "intellectual property" is an entirely artificial construct that goes against the nature of information and knowledge, forcing information to behave like physical goods, so it's not unfair to the creators in an economy that's built around trading physical goods. IP laws were built on moral arguments, so it's only fair to change them on moral grounds too.) -- [0] - Greed is more visible in the LLM theatre of this conflict, because with textual content there's vastly more people who believe that they're entitled to compensation just because some comments they wrote on the Internet may have been part of the training dataset, and are appalled to see LLM providers get paid for the service while they are not. This Dog in the Manger mentality is distinct from that of people whose output was used in training a model that now directly competes with them for their job; the latter have legitimate ethical reasons to complain. [1] - Even though myself I am for treating training datasets to generative AI as exempt from copyright. I think it'll be better for society in general - but I recognize it's easy for me to say it, because I'm not the one being rugpulled out of a career path by GenAI, watching it going from 0 to being half of the way towards automating away visual arts, in just ~5 years.
- skissane 2y ago
  > they have a legitimate argument that this is an unexpected, undeserved, unfair calamity for them, threatening to derail their lives, and lives of their dependents, across the entire sector - and therefore that laws should be changed to shield them, or compensate them for the loss. After all, that's what laws are for. Lots of people have had their lives disrupted by technological and economic changes before - entire careers which existed a century ago are now gone. Given society provided little or no compensation for prior such cases of disruption, what’s the argument for doing differently here?
  - TeMPOraL 2y ago
    Moral growth and learning from history?
    - skissane 2y ago
      There’s a big risk that you end up creating a scheme to compensate for technological disruption in one industry and then fail to do so in another, based on the political clout / mindshare / media attention each has - and then there are many people in even worse personal situations (through no fault of their own) who would also miss out. Wouldn’t a better alternative be to work on improving social safety nets for everybody, as opposed to providing a bespoke one for a single industry?
      - TeMPOraL 2y ago
        > Wouldn’t a better alternative be to work on improving social safety nets for everybody, as opposed to providing a bespoke one for a single industry? Yes, but: 1) It's not really an exclusive choice; different people can pursue different angles, including all of them - one can both seek immediate support/compensation for the specific case they're the victim of and seek longer-term solution for everyone who'd face the same problem in the future. 2) A bespoke solution is much more likely to be achievable than a general one. 3) I don't believe it would be good for society for artists to succeed in curtailing generative AI! But, should they succeed, I imagine the consequences will encourage people to seek the more general solution that mitigates occupational damage of GenAI while preserving its availability, instead of having to deal with a series of bespoke stopgaps that also kills GenAI entirely. 4) Not that banning GenAI has any chance of succeeding - the most we'd get is it being unavailable in some countries, who'd then be at a disadvantage in competition with countries that embraced it. Again, I'm not in favor of banning GenAI - on the contrary, I'm in favor of giving a blanket exception from copyright laws for purposes of training generative models. However, I recognize the plight of artists and other people who are feeling the negative economic impact on their jobs right now (and hell, my own line of work - software development - is still one of the most at risk in the near to mid-term, too); I wish for a solution that will help them (and others about to be in this situation), but in the meantime, I don't begrudge them for trying to fight it - I think they have full right to. I only have problems with people who oppose AI because they feel that Big AI is depriving them of opportunity to seek rent from society for the value AI models are creating.
skywhopper 2y ago
In the context of encouraging art, it totally is! Copyright and patents are 100% artificial and invented legal concepts that are based solely on the distaste for others profiting off a creator’s ideas. The reason for them is to encourage creativity by allowing creators to profit off new ideas. So there’s no reason why “distaste” about AI abuse of human artists’ work shouldn’t be a valid reason to regulate or ban it. If society values the creation of new art and inventions, then it will create artificial barriers to encourage their creation.
- bmacho 2y ago
  Yup, banning AI for the sake of artist would be exactly the same as the current copyright laws. (Also they are attacking AI not purely for fear of their jobs, but bc it is illegal already.)
- _bin_ 2y ago
  Disagree. Authority is given Congress to establish an IP regime for the purpose of "promot[ing] the progress of science and useful arts". You would have to justify how banning gen AI is a. feasible at all, particularly with open-weight models; and b. how it "promotes the progress of useful arts." You would lose in court because it's very difficult to argue that keeping art as a skilled craftsman's trade is worse for its progress than lowering the barriers to individuals expressing what they see. I think bad AI makes bad output and so a few people are worried it will replace good human art with bad AI art. Realistically, the stuff it's replacing now is bad human art: stock photos and clipart stuff that weren't really creative expression to start with. As it improves, we'll be increasingly able to go do a targeted inpaint to create images that more closely match our creative vision. There's a path here that lowers the barriers for someone getting his ideas into a visual form and that's an unambiguous good, unless you're one of the "craftsmen" who invested time to learn the old way. It's almost exactly the same as AI development. As an experienced dev who knows the ins and outs really well I look at AI code and say, "wow, that's garbage." But people are using it to make unimportant webshit frontends, not do "serious work". Once it can do "serious work" that will decrease the number of jobs in the field but be good for software development as a whole.
cratermoon 2y ago
> Right, but "distaste" isn't grounds for trying to ban something I disagree. There are many laws on the books codifying social distastes. They keep your local vice squad busy.
- _bin_ 2y ago
  I thought most people supported moving away from that and towards a more socially liberal model. If we're no longer doing that I have a whole stack of socially conservative policies I guess I'll go back to pushing. I don't think y'all really want to go down this road; it leads straight back to the nineties republicans holding senate hearings on what's acceptable content for a music album.
  - fc417fc802 2y ago
    Many laws come down to distaste at the root. There's usually an alternative angle about market efficiency or social stability or whatever if you want to frame it that way. The same applies in this case as well. For but a few examples consider laws regarding gambling, many aspects of zoning, or deceptive marketing. What's the purpose of the law if not providing stability? Why should social issues be exempted from that?
computerthings 2y ago
[dead]
weregiraffe 2y ago
>Right, but "distaste" isn't grounds for trying to ban something. https://en.wikipedia.org/wiki/United_States_obscenity_law https://en.wikipedia.org/wiki/United_States_obscenity_law

Riverheart 2y ago

“A legal reality where you can only train AI on content you've licensed would be the worst for everybody bar massive companies, legacy artists included.” Care to elaborate? Also, saying artists only concern themselves with the legality of art used in AI because of distaste when there are legal cases where their art has been appropriated seems like a bold position to take. It’s a practice founded on scooping everything up without care for origin or attribution and it’s not like it’s a transparent process. There are people that literally go out of their way to let artists know they’re training on their art and taunt them about it online. Is it unusual they would assume bad faith from those purporting to train their AI legally when participation up till now has either been involuntary or opt out? Rolling out AI features when your customers are artists is tone deaf at best and trolling at worst.

Workaccount2 2y ago
There is no "scooping up", the models aren't massive archives of copied art. People either don't understand how these models work or they purposely misrepresent it (or purposely refuse to understand it). Showing the model an picture doesn't create a copy of that picture in it's "brain". It moves a bunch of vectors around that captures an "essence" of what the image is. The next image shown from a totally different artist with a totally different style may well move around many of those same vectors again. But suffice to say, there is no copy of the picture anywhere inside of it. This also why these models hallucinate so much, they are not drawing from a bank of copies, they are working off of a fuzzy memory.
- TeMPOraL 2y ago
  > People either don't understand how these models work or they purposely misrepresent it (or purposely refuse to understand it). Not only that, they also assume or pretend that this is obviously violating copyright, when in fact this is a) not clear, and b) pending determination by courts and legislators around the world. FWIW, I agree with your perspective on training, but I also accept that artists have legitimate moral grounds to complain and try to fight it - so I don't really like to argue about this with them; my pet peeve is on the LLM side of things, where the loudest arguments come from people who are envious and feel entitled, even though they have no personal stake in this.

spoaceman7777 2y ago

> Adobe isn't trying to be ethical, they are trying to be more legally compliant Is the implication of this statement that using AI for image editing and creation is inherently unethical? Is that really how people feel?

mtndew4brkfst 2y ago
For creation, yes, because of the provenance of the training data that got us here. It was acquired unethically in the overwhelming majority of cases. Using models derived from that training is laundering and anonymizing the existing creativity of other humans and then still staking the claim "I made this", like the stick figure comic. It's ghoulish.
- skissane 2y ago
  There exist image generation models that were trained on purely licensed content, e.g. Getty’s. I don’t know about Adobe’s specifically-but if not, it seems like a problem Adobe could easily fix-either buy/license a stock image library for AI training (maybe they already have one), and use that to train their own model-or else license someone else’s model e.g. Getty’s
  - spookie 2y ago
    Well they do license the art they use, but in... let's say... "interesting" ways through their ToS.
  - bolognafairy 2y ago
    They are training using licensed images! That’s the thing! There’s some sort of ridiculous brainworm infecting certain online groups that has them believing that stealing content is inherent in using generative AI. I watch this all quite closely, and It’s chronically online, anime / fursona profile picture, artists. Exact same thing happened when that ‘open’ trust and safety platform was announced a few months ago, which used “AI” in its marketing material. This exact same group of people—not even remotely the target audience for this B2B T&S product—absolutely lost it on Bluesky. “We don’t want AI everywhere!” “You’re taking the humanity out of everything!” “This is so unethical!” When you tell them that machine learning has been used in content moderation for decades, they won’t have a bar of it. Nor when you explain that T&S AI isn’t generative and almost certainly isn’t using “stolen” data. I had countless people legitimately say that having humans have to sift through gore and CSAM is a Good Thing because it gives them jobs, which AI is taking away. It’s all the same sort of online presence. Anime profile picture, Ko-fi in bio, “minors dni”, talking about not getting “commissions” anymore. It genuinely feels like a psy-op / false flag operation or something.

no_wizard 2y ago

> A legal reality where you can only train AI on content you've licensed would be the worst for everybody bar massive companies, legacy artists included. Quite an assertion. Why exactly would this be true?

drilbo 2y ago
who else has would ever have a significantly large store of licensed material?
- fc417fc802 2y ago
  Or alternatively, who else could afford the licensing costs?

dinkumthinkum 2y ago

I'm curious why you think it would be worse for everybody? This argument seems to depend on the assumption that if something makes AI less viable then the situation for human beings is worse overall. I don't think many actual people would accept that premise.

crimony 2y ago
It's worse only if AI turns out to be of high value. In that case only large companies that can afford to license training data will be dominant.

tbrownaw 2y ago

> Adobe isn't trying to be ethical, they are trying to be more legally compliant, Ethics (as opposed to morals) is about codified rules. The law is a set of codified rules. So are these really that different (beyond how the law is a hodge-podge and usually a minimum requirement rather than an ideal to reach for)?

zmmmmm 2y ago

The ship has sailed, but I can understand artists feeling that no matter how any AI is trained prospectively, it was only made possible because the methods to do so were learned through unethical means - we now know the exact model architectures, efficient training methods and types of training data needed so that companies like Adobe can recreate it with a fraction of the cost. We obviously can never unscramble that egg, which is sad because it probably means there will never be a way to make such people feel OK about AI.