OpenAI Defends Use of Copyrighted Material to Train AI Models
-
OpenAI says AI models like ChatGPT would be "impossible" without using copyrighted material in training data.
-
Models gain abilities from large-scale training on content scraped from the internet without permission.
-
OpenAI claims virtually all human expression is copyrighted today, limiting training data.
-
OpenAI asserts training AI models this way is fair use under long-standing legal precedents.
-
OpenAI previously defended its AI training data use as fair use in response to other copyright lawsuits.