CREATORS UNITE TO CALL FOR LEGAL PROTECTIONS AGAINST AI INFRINGEMENT

CREATORS UNITE TO CALL FOR LEGAL PROTECTIONS AGAINST AI INFRINGEMENT

Individual artists have complained, protested, and even sued OpenAI, Meta, and other AI developers for using their work to train AIs without compensating creators or asking their permission. 

Much of that work is copyrighted by its human creators and AI developers have used it to train their AIs without asking permission from, or offering payment to, the copyright owners.

Now the creative community has taken the next step: several arts organizations representing more than six million artists and performers around the world have banded together to propose a sort of “creators’ bill of rights” related to AI:  

  1. Creators must have the legal right to authorize or deny AI developers the use of their material and have the right to be compensated for giving their permission.
  2. Creators have a right to know how their work will be used in training an AI.
  3. Exceptions to the above two principles must be severely limited—no loopholes.
  4. Creators must receive credit in the AI for allowing their work to be used.
  5. AI systems must be required to disclose which materials it offers up are copyrighted and which are generated by AI.
  6. AI companies must keep records of their use of copyrighted materials and be subject to audits.
  7. AI-generated work should not be afforded the same copyright and other legal protections as work created by humans.
  8. Fully autonomous AI works should not receive the same legal protections as human creations.

TRENDPOST: AI developers are hoping to negate these demands by “synthesizing” the data on which the next generation of AI is trained. Instead of using original sources on the Internet, new AIs would be infused with the homogenized data from previous AIs. That would remove a significant amount of identifiably original content, similar to the way you can’t taste the individual ingredients in a cake mix when they all go into a batter.

However, that doesn’t mean that the flour and sugar aren’t there; the original AIs used to train new ones would still be passing on content that, in some way, depends on creators’ works.

Ultimately, the battle is likely to be fought out in court.

Skip to content