Blog Posts

In a world of deepfakes, AI-generated robocalls, and fabricated media, the internet has no native trust layer. This paper proposes building one. Starting from the observation that social media algorithms were designed to maximize engagement — not truth — it argues for a layered "trust stack": a provenance layer using cryptographic signatures to verify content origin and editing history; a semantic layer using AI to assess the likelihood that claims are accurate; and a user-facing signal as intuitive as a browser padlock. Drawing on behavioral science and analogies from HTTPS adoption, nutrition labeling, and Creative Commons licensing, the paper outlines how a Trust-as-a-Service infrastructure could be built, funded, and deployed — and why doing so is urgent.

Examines the role of the human prompt engineer in AI-generated content and asks whether prompt engineering constitutes the kind of creative authorship that copyright law requires. Argues that prompts are typically ideas rather than protectable expression, and that the creative choices embodied in most AI outputs trace back to the machine rather than to the human who instructed it — with significant consequences for how copyright law should respond to generative AI.

Introduces the metaphor of "self-driving culture" to capture what is at stake when AI systems produce the literary works, journalism, and music through which societies understand and reshape themselves. Argues that delegating cultural production to machines is not merely an economic disruption but a civilizational shift — one that IP law, still built around the assumption of human authorship, is not yet equipped to address.

Considers the broader implications of generative AI for human work and creativity — what happens when machines can perform not just physical but cognitive and creative tasks better, faster, and more cheaply than humans. Frames the challenge as an inversion of the historical relationship between human labor and technological tools, and asks what law and policy can do to preserve meaningful space for human contribution.

A practical overview of the principal intellectual property questions raised by generative AI, organized as a checklist for practitioners, policymakers, and scholars. Covers copyright authorship and originality, patent inventorship, training data and fair use, ownership of outputs, and liability for infringement — providing a structured map of an area in which doctrine is still rapidly evolving.