The AI industry relies on the 'fair use' doctrine, which permits the use of copyrighted material without permission in certain circumstances. However, this gives the industry the ability to exploit copyrighted works without compensating authors while humans face restrictions, argues Ben Sobel.

The author criticises 'free culture' activists for supporting AI companies, pointing out that their opposition to copyright expansion unintentionally supports the AI industry's defence. Instead, the author suggests enforcing copyright regulations equally on AI learners and humans to highlight the law's disregard for learning and potentially lead to relaxed regulations for humans.

The concept of 'copyright accelerationism' is introduced as a tactic to challenge legal rules that hinder ordinary people's access to creative works. The author claims that copyright protections have limited access to cultural works over the past 50 years, benefiting only a small group of rights holders.

The author questions the AI industry's reliance on the fair-use doctrine and argues that a similar argument made by a human would not hold up in court. To address this issue, the author proposes that decelerationists, who seek to halt copyright expansion, support reforms that eliminate copyright doctrines inhibiting both human and machine learning. This, according to the author, would challenge the harmful aspects of copyright law and create a more equitable copyright system.