Abstract:
Generative AI has revealed a blind spot in liberal private law, of which intellectual property law is illustrative. Copyright’s statutory verbs fail to describe what models actually do with training data, while contract and property doctrines have enabled enclosure. The result is a political economy of double extraction: Web 2.0’s attention harvest and model developers’ appropriation of human traces to generate synthetic substitutes. The rise of “AI slop” makes this approach unsustainable, as authentic human contributions become scarce and newly valuable. This paper argues that liberal private law can be refitted for the digital economy, but the challenge is no longer doctrinal fine-tuning.