3 ms·

So it makes generative AI essentially unusable, because you don't know if the output is plagiarism or not, so you'd just doubt it always and never use it.

Too 3y ago
It’s usable for internal content, maybe even a small public blog where you sprinkle in some generated pictures instead of stock photos. Nobody will care if your school project contains a Mario holding a Coca Cola. It’s once you start monetizing and publishing on bigger scale, without appropriating, it gets interesting.
- TheRoque 3y ago
  The thing is, this market is way too small.
Art9681 3y ago
The same tools and methods used to detect plagiarism or copyright violation can be employed to check the generated content and modify it just enough to fall outside the scope of any law banning its use for profit. Inevitably, a platform will emerge to do this. From a technical standpoint it is game over. This is indisputable. By the end of next year many models and software tools will exist whose entire purpose will be to do just this. And the ones deploying those tools at scale will be businesses like the New York Times having realized that the only way to survive this is to float with the unstopable tide. Nothing short of absolute privacy violation will stop web unauthorized web scraping. Tools exist today that automate a browser and easily fool the web servers into thinking its just a person clicking around. It works quite well. It works with authorized accounts. It works in the same way any person would visit a site, highlight some text and copy it. What are they going to do? Require the end user's web cam to be on so they can verify a human is navigating next? Its game over folks. And this is going to happen with or without our approval and any government that limits the potential use of this is only giving nations that dont a large economic advantage. Interesting times ahead.
- blizzard_dev_17 3y ago
  > What are they going to do? Litigation. Hiding behind changing geolocation won't do much since law enforcement has access to the same tools but the difference is they can force companies (your ISP, google) to comply. Obsolete take, next!
koliber 3y ago
No. It”s still very helpful. However you can not blindly take whatever it produces and publish it. Sometimes it hallucinates. Sometimes it draws weird looking hands. Sometimes it generates copyrighted materials. Check the work it produces.
- schmichael 3y ago
  Check it against what exactly? How do you, the end user, determine an image does not infringe?
- TheRoque 3y ago
  Then the generative tools should just give the sources of the inspiration of the AI and make them aware of what they are using, instead of saying "nope, not my problem".
  - koliber 3y ago
    The consumer will be free to choose what they demand from their tooling. If consumers decide that they only want to use generative AI that does what you propose, they’ll vote with their wallet. If they decide to use other ways of checking for IP infringement, they will. If they choose to ignore the issue, IP owners will bring up violations, like the NYT did. “Buyer beware” has been a motto since ancient times.

Too 3y ago

It’s usable for internal content, maybe even a small public blog where you sprinkle in some generated pictures instead of stock photos. Nobody will care if your school project contains a Mario holding a Coca Cola. It’s once you start monetizing and publishing on bigger scale, without appropriating, it gets interesting.

TheRoque 3y ago
The thing is, this market is way too small.

Art9681 3y ago

The same tools and methods used to detect plagiarism or copyright violation can be employed to check the generated content and modify it just enough to fall outside the scope of any law banning its use for profit. Inevitably, a platform will emerge to do this. From a technical standpoint it is game over. This is indisputable. By the end of next year many models and software tools will exist whose entire purpose will be to do just this. And the ones deploying those tools at scale will be businesses like the New York Times having realized that the only way to survive this is to float with the unstopable tide. Nothing short of absolute privacy violation will stop web unauthorized web scraping. Tools exist today that automate a browser and easily fool the web servers into thinking its just a person clicking around. It works quite well. It works with authorized accounts. It works in the same way any person would visit a site, highlight some text and copy it. What are they going to do? Require the end user's web cam to be on so they can verify a human is navigating next? Its game over folks. And this is going to happen with or without our approval and any government that limits the potential use of this is only giving nations that dont a large economic advantage. Interesting times ahead.

blizzard_dev_17 3y ago
> What are they going to do? Litigation. Hiding behind changing geolocation won't do much since law enforcement has access to the same tools but the difference is they can force companies (your ISP, google) to comply. Obsolete take, next!

koliber 3y ago

No. It”s still very helpful. However you can not blindly take whatever it produces and publish it. Sometimes it hallucinates. Sometimes it draws weird looking hands. Sometimes it generates copyrighted materials. Check the work it produces.

schmichael 3y ago
Check it against what exactly? How do you, the end user, determine an image does not infringe?
TheRoque 3y ago
Then the generative tools should just give the sources of the inspiration of the AI and make them aware of what they are using, instead of saying "nope, not my problem".
- koliber 3y ago
  The consumer will be free to choose what they demand from their tooling. If consumers decide that they only want to use generative AI that does what you propose, they’ll vote with their wallet. If they decide to use other ways of checking for IP infringement, they will. If they choose to ignore the issue, IP owners will bring up violations, like the NYT did. “Buyer beware” has been a motto since ancient times.