AI as an Inventing Tool: AI’s Challenges to Patent Law – Inventorship (Panel 2)
Why It Matters
Because AI-generated inventions strain existing inventorship rules, policymakers must adapt patent frameworks to preserve innovation incentives and ensure transparent ownership.
Key Takeaways
- •Patent law mandates a natural-person inventor under Paris Convention.
- •AI-generated inventions challenge traditional inventorship definitions and enforcement.
- •USPTO currently adopts a "don't ask, don't tell" stance on AI contributions.
- •Federal Circuit cases (Thaler v. Vidal) reject AI as legal inventor.
- •Industry consensus leans toward treating AI output as human-directed invention.
Summary
The panel examined how AI tools upend the traditional notion of inventorship in patent law, focusing on the requirement that a natural person be named as inventor under the Paris Convention and U.S. statutes.
Speakers highlighted that U.S. law, unlike copyright’s work-for-hire doctrine, still demands at least one human inventor, while the USPTO's current "don't ask, don't tell" approach sidesteps rigorous scrutiny of AI contributions. They noted the 2011 America Invents Act’s reduced penalties for incorrect inventorship and contrasted stricter European standards that have already invalidated patents for missing inventors.
Professor Crouch likened AI-driven design to a mouse generating antibodies, arguing that the mouse functions as a prompt-engineer. The Federal Circuit’s Thaler v. Vidal decision was cited, where the USPTO rejected "DAB Dabbas" as an AI-named inventor, reaffirming the statutory language that invention must be conceived in a human mind.
The discussion underscored the looming need for legislative or regulatory reforms to accommodate AI-generated inventions, lest companies resort to secrecy and lose the public-policy benefits of disclosure, while patent owners grapple with ownership and enforcement uncertainties.
Comments
Want to join the conversation?
Loading comments...