Patents after the Invention Machine

Abstract

While contemporary legal and media discourse surrounding artificial intelligence (AI) remains predominantly focused on the doctrinal question of machine inventorship, this paper argues that such debates misidentify the primary challenge confronting the intellectual property regime. From an economic perspective, the core disruption of the "invention machine" lies in its capacity to drastically reduce the marginal cost of generating candidate inventions. When the production of novel ideas becomes cheap and scalable, traditional patent protections risk functioning as distortionary windfalls rather than necessary market incentives, potentially granting broad exclusionary rights over outputs that would have emerged absent legal protection. Consequently, this paper reframes the policy debate toward a fundamental economic inquiry: whether society should incentivize AI-generated inventions through the patent system at all, and if so, how the legal framework must adapt to preserve incentives in capital-intensive industries—such as pharmaceuticals and biotechnology—where the transition from conceptualization to commercialization remains highly costly.

Keywords: artifical inventor, artifical intelligence, patent