The issue is simply reproduction of original works.
Plenty of people mimic the style of other artists. They do this by studying the style of the artist they intend to mimic. Why is it different when a machine does the same thing?
OpenAI should not be held back from subscribing to a research publication, or buying college textbooks, etc. As long as the original works are not reproduced and the underlying concepts are applied, there are no intellectual property issues. You can’t even say the commercial application of the text is the issue, because I can go to school and use my knowledge to start a company.
I understand that in some select scenarios, ChatGPT has been tricked into outputting training data. Seems to me they should focus on fixing that, as it would avoid IP issues moving forward.
AI image creation tools are apparently both artistically empty, incapable of creating anything artistically interesting, and also a existential threat to visual artists. Hmm, wonder what this says about the artistic merits of the work of furry porn commission artist #7302.
Retail workers can be replaced with self checkout, translators can be replaced with machine translation, auto workers can be replaced with robotic arms, specialist machinists can be replaced with CNC mills. But illustrators must be where we draw the line.
It’s different because a machine can be replicated and can produce results at a rate that hundreds of humans can’t match. If a human wants to replicate your art style, they have to invest a lot of time into learning art and practicing your style. A machine doesn’t have to do these things.
This would be fine if we weren’t living in a capitalist society, but since we do, this will only result in further transfer of assets towards the rich.
The issue is simply reproduction of original works.
Plenty of people mimic the style of other artists. They do this by studying the style of the artist they intend to mimic. Why is it different when a machine does the same thing?
No, the issue is commercial use of copirighted material as data to train the models.
It’s not. People are just afraid of being replaced, especially when they weren’t that original or creative in the first place.
Honestly, it extends beyond creative works.
OpenAI should not be held back from subscribing to a research publication, or buying college textbooks, etc. As long as the original works are not reproduced and the underlying concepts are applied, there are no intellectual property issues. You can’t even say the commercial application of the text is the issue, because I can go to school and use my knowledge to start a company.
I understand that in some select scenarios, ChatGPT has been tricked into outputting training data. Seems to me they should focus on fixing that, as it would avoid IP issues moving forward.
AI image creation tools are apparently both artistically empty, incapable of creating anything artistically interesting, and also a existential threat to visual artists. Hmm, wonder what this says about the artistic merits of the work of furry porn commission artist #7302.
Retail workers can be replaced with self checkout, translators can be replaced with machine translation, auto workers can be replaced with robotic arms, specialist machinists can be replaced with CNC mills. But illustrators must be where we draw the line.
It’s different because a machine can be replicated and can produce results at a rate that hundreds of humans can’t match. If a human wants to replicate your art style, they have to invest a lot of time into learning art and practicing your style. A machine doesn’t have to do these things.
This would be fine if we weren’t living in a capitalist society, but since we do, this will only result in further transfer of assets towards the rich.