AI and the Value of Human Authorship: The Fight for Copyright
The training data for generative AI systems contains millions upon millions of copyrighted works. The rights holders of this material were neither asked if they consented to this form of use nor were they compensated for it. Many celebrate the fact that texts, images, and musical works are publicly available on the internet as an achievement for free education and participation in humanity’s cultural wealth. Of course, no one thought about the for-profit companies that develop large language models on this basis.
