AI is stealing writers’ words and jobs…

[Jfdlsjfd]

Eventually it will all come out. The question to be answered is whether the dissemination and use of Generative AI is to the net public benefit which is at the heart of the exception granted news agencies, and will be necessary for AI to have the same exception granted. I'm dubious on that.

Honestly, the research exception granted for non-profit entities as part of the TDM exception and the opt-out rule even for commercial purpose seems a big move toward this. I am not as dubious as you are over tis outcome. Several studies, notably by PwC, estimated (or wildly guessed, I am dubious of precise numbers given the methodology) the economic windfall of generative AI was in the range of trillions USD each year. Making it possible by disseminating access to AI seems a very good public benefit (increasing the taxable wealth). It is even more tempting if major countries start banning it, as the competition for hosting these industries will narrow, leading to a concentration of the taxable wealth in fewer countries, who'll be incentivized to provide this exception.

 

[trappedslider]

No, it was done in a different thread by someone else, I'm not sure what the prompt was.

I don't play with ChatGPT, just Dall-E to see how far we can take concept generation without paying anyone.

(Answer: Very far.)

Yeah it was done with one of the others but they intentionally made the prompt very specific in what they wanted because when I tried a prompt based on what they wanted and got told "No, I can't do that" by ChatGPT they i guess weren't happy. That's why I also ask "what prompt did you use?"

Eventually it will all come out. The question to be answered is whether the dissemination and use of Generative AI is to the net public benefit which is at the heart of the exception granted news agencies, and will be necessary for AI to have the same exception granted. I'm dubious on that. It's also one of the reasons that I don't put this issue in the same category as wagon wheel makers or ice delivery men.

I wonder if they will take "intent" into account, but Open AI has placed the burden on the user much like other companies have done with the tools they make.

 

[Scribe]

Yeah it was done with one of the others but they intentionally made the prompt very specific in what they wanted because when I tried a prompt based on what they wanted and got told "No, I can't do that" by ChatGPT they i guess weren't happy. That's why I also ask "what prompt did you use?"

Which, is obviously a lie. It could do that, but it is prevented by explicit direction, to prevent from being sued.

So, did these models train by ripping out of protected texts, which the companies pirated?

Obviously.

 

[Ryujin]

Yeah it was done with one of the others but they intentionally made the prompt very specific in what they wanted because when I tried a prompt based on what they wanted and got told "No, I can't do that" by ChatGPT they i guess weren't happy. That's why I also ask "what prompt did you use?"

I wonder if they will take "intent" into account, but Open AI has placed the burden on the user much like other companies have done with the tools they make.

The same rationale was used by gun manufacturers and cigarette companies. Kind of works for the former. Ultimately failed miserably for the latter. To take it to the extreme, "I didn't plagiarize anything. I just created the tools that make it trivial for other people to do so." At that level it's right up there with Torrent creators blaming the users for getting stuff via bittorrent.

 

[Ryujin]

Honestly, the research exception granted for non-profit entities as part of the TDM exception and the opt-out rule even for commercial purpose seems a big move toward this. I am not as dubious as you are over tis outcome. Several studies, notably by PwC, estimated (or wildly guessed, I am dubious of precise numbers given the methodology) the economic windfall of generative AI was in the range of trillions USD each year. Making it possible by disseminating access to AI seems a very good public benefit (increasing the taxable wealth). It is even more tempting if major countries start banning it, as the competition for hosting these industries will narrow, leading to a concentration of the taxable wealth in fewer countries, who'll be incentivized to provide this exception.

I see it differently. It's more about wealth concentration but to go into that would immediately run afoul of the rules, so I'll stop with just the mention.

 

[trappedslider]

Which, is obviously a lie. It could do that, but it is prevented by explicit direction, to prevent from being sued.

So, did these models train by ripping out of protected texts, which the companies pirated?

Obviously.

Honestly, at this point I feel that it's moved beyond the issue of "how" and "what they used for training" (old joke, every book is a remix of the dictionary) and has moved into "what do you plan to do with this?" hence why I'm all about intent. I know the intention of the companies (make money derp)

Obviously others feel that it should just stop at the issues of "how they were trained" and/or "It never should have happened to begin with"

The same rationale was used by gun manufacturers and cigarette companies. Kind of works for the former. Ultimately failed miserably for the latter. To take it to the extreme, "I didn't plagiarize anything. I just created the tools that make it trivial for other people to do so." At that level it's right up there with Torrent creators blaming the users for getting stuff via bittorrent.

cigarettes' are still made (I'm not going to touch guns cuz politics) but they are heavily regulated. I also think that in this day and age Fair use law needs to be reworked (it should have been years ago but that's beside the point)

I see it differently. It's more about wealth concentration but to go into that would immediately run afoul of the rules, so I'll stop with just the mention.

And I wish to publicly thank you for taking my offer of taking that part to PM now I know what your view is and all that .

 

[Scribe]

Honestly, at this point I feel that it's moved beyond the issue of "how" and "what they used for training" (old joke, every book is a remix of the dictionary) and has moved into "what do you plan to do with this?" hence why I'm all about intent. I know the intention of the companies (make money derp)

I mean I'm kind of with you.

How? Stealing.
What they used? Everything they could touch.
What for? Money.

I dont find any of it really all that deep. Thats it. They intend to make as much money as possible, otherwise billions would not be being invested.

Its not going to bring about peace on earth. We are a bunch of apes with a dangerously inflated sense of self importance and self righteousness. AI wont fix that.

 

[Jfdlsjfd]

I see it differently. It's more about wealth concentration but to go into that would immediately run afoul of the rules, so I'll stop with just the mention.

I think it's still within the rules to say that SOME PEOPLE (at least, Forbes editors, who create a ranking of countries by numbers of ultra-high net worth individuals, favoring of course low-Gini countries) would see wealth concentration by itself as being a public benefit. It is possible that the appreciation of the public benefit would also be determined by past ability/will of their society to turn GDP increase into a balanced improvement for its population, so countries would have another metric to determine net public interest, leading to different choices over this matter in different jurisdictions.

I can't see research centres relocating to Nepal or Vanuatu, but I can see them going to Europe, Japan, Singapore, China... whichever is the more attractive after they have settled their stance on AI.

 

[overgeeked]

I mean I'm kind of with you.

How? Stealing.
What they used? Everything they could touch.
What for? Money.

I dont find any of it really all that deep. Thats it. They intend to make as much money as possible, otherwise billions would not be being invested.

Its not going to bring about peace on earth. We are a bunch of apes with a dangerously inflated sense of self importance and self righteousness. AI wont fix that.

Exactly. It’s not some deep conspiracy or fifth-dimensional chess. Corporations gambled they could get away with stealing the IP of billions of people long enough to make billions of dollars. So far, they seem to be right.

 

[trappedslider]

I mean I'm kind of with you.

How? Stealing.
What they used? Everything they could touch.
What for? Money.

I dont find any of it really all that deep. Thats it. They intend to make as much money as possible, otherwise billions would not be being invested.

Its not going to bring about peace on earth. We are a bunch of apes with a dangerously inflated sense of self importance and self righteousness. AI wont fix that.

That's why I will happily go yeah they (may in some cases instead of fair use law) have broke copy right law to make the thing, but I wanna know what you're doing with it. Here's what I'm doing with it. I find the former discussion to be boring, while the other discussion is more fun. I'm sorry if that doesn't go with your ( general your) axe grinding but whatever.

I posted over in https://www.enworld.org/threads/i-used-chatgp-for-my-setting-and-it-was-awesome.698230/ last summer, about how i used the then edition of chatgpt to help flush out the setting and I've recently been using it to help fix errors in my writing (spelling, grammar continuity word sense)

As for the art stuff like what nightcafe offers, I've just been messing around with it while still commissioning pieces from actual artist.