Today, Explained AI is killing the internet
145 snips
Jul 30, 2025 
Jason Kebler, a tech reporter and co-founder of 404 Media, dives into the tumultuous intersection of AI and copyright law. He discusses landmark lawsuits affecting authors and highlights the ethical dilemmas of using unlicensed materials for AI training. Kebler addresses the penguins’ impact of AI on Google Search, emphasizing concerns over content quality and shifts in user experience. He also explores how these developments threaten online creativity, while public interactions evolve amidst growing distrust towards AI.
AI Snips
Chapters
Transcript
Episode notes
AI Training Combines Fair Use and Piracy
- AI companies train models on massive copyrighted data sets scraped from piracy sites and other sources.
- Courts may rule training use fair, but acquiring data illegally remains infringement.
Legal Nuances of AI Copyright Use
- The court ruled AI training on scraped copyrighted works is transformative fair use.
- Nonetheless, illegal acquisition methods are not protected, creating a complex legal landscape.
Anthropic's Dual Data Acquisition Approach
- Anthropic initially pirated books from torrent and piracy sites to expedite data acquisition.
- Later, they bought physical books from used stores, scanned them to train their AI legitimately.