Should AI-Generated Work Be Eligible for Copyright?

• The logic of copyright is not about rewarding human creative suffering — it is about incentivizing the production of creative works for public benefit. If AI systems produce valuable creative output, denying copyright creates a collective-action problem: anyone can copy the output instantly, so no company can recoup the cost of building the AI system. The result is less creative AI, not more human creativity.
• The 'human author' requirement is a policy choice, not a natural law. Corporate authorship was invented in the 19th century to allow publishers and studios to hold copyright — not because corporations are creative, but because the law needed to incentivize creative investment. The same logic applies to AI: assign copyright to the company that built the model and directed the creative output.
• Current US Copyright Office guidance creates an absurd gradient: a human who types one sentence directing an AI to generate 10,000 words holds copyright on the whole; a human who generates the same work with slightly less direction holds nothing. This is a doctrinal incoherence that requires a principled framework — and the most principled framework is calibrated to the degree of human creative input, not a binary on/off switch.

Debater: To be announced

• The Supreme Court in Thaler v. Vidal (2023) and the Copyright Office's Zarya of the Dawn decision both held that copyright requires human authorship — and both were right. The Constitution grants Congress power to protect creative work to promote the arts and sciences. AI companies are not the kind of entity that responds to copyright incentives the way human authors do; they are already building AI regardless.
• AI systems are trained on billions of existing human creative works, most without explicit licensing. Granting AI companies copyright over the output doubles the injustice: first, they extract value from human creators without compensation; then they monopolize the derivative outputs those creators never licensed. The correct policy is to resolve the training-data question, not to reward the extraction with new ownership rights.
• If AI-generated work enters the public domain, it enriches the commons rather than concentrating ownership in AI companies. Musicians, filmmakers, and writers could use AI output as raw material for genuinely human-authored creative work. A public-domain AI commons is more consistent with copyright's goal — public benefit — than a regime that gives Disney, Google, and OpenAI perpetual monopolies over AI-generated content.

Debater: To be announced