The United States Patent and Trademark Office (USPTO) has issued FR Doc. 2025-21457, a revised framework for determining inventorship in the era of artificial intelligence. This new guidance fully replaces and revokes the USPTO’s inventorship guidance of February 13, 2024, and clarifies that, under current U.S. law, only natural persons may be named as inventors while AI systems are treated as tools rather than collaborators.

Why This New Document Is Important

FR Doc. 2025-21457 is more than an update: it replaces and revokes the USPTO’s February 13, 2024 inventorship guidance, providing a clearer, more streamlined framework for assessing patent applications involving AI.

Its importance lies in three key areas:

1. It Reinforces Legal Certainty in a Fast-Moving Technological Landscape

The document reaffirms a foundational principle of U.S. patent law:
Only natural persons can be inventors.
AI systems, no matter how autonomous or sophisticated, cannot be named as inventors or co-inventors.

This clarity is essential for companies, researchers, and policymakers navigating the growing integration of AI in R&D.

2. It Reasserts a Traditional, Stable Inventorship Standard

Instead of introducing new tests for AI-assisted invention, the USPTO returns to the longstanding definition of “conception”:

the formation in the mind of an inventor of a definite and permanent idea of the complete invention.

By explicitly reaffirming the traditional standard, FR Doc. 2025-21457 ensures predictability and continuity, avoiding legal uncertainty as AI advances.

3. It Treats AI as a Tool—Not a Collaborator

The new guidance draws a firm line: AI may assist, accelerate, or generate outputs, but it does so as an instrument under human direction.

This distinction preserves the human-centered structure of the patent system while allowing innovation to continue integrating AI tools.

---

How the New Guidance Differs from the February 13, 2024 Version

While the 2024 guidance was an important early step, FR Doc. 2025-21457 introduces three major shifts:

1. The 2024 guidance introduced a “significant human contribution” test.

The 2025 guidance removes it.

The February 13, 2024 document emphasized that a human must make a “significant” contribution to the invention—a concept drawn from joint inventorship case law.

The 2025 version discards this emphasis and focuses instead on whether the human meets the traditional legal definition of conception, without the additional threshold of «significance.»

2. Clearer, simpler criteria replace the more layered 2024 framework

The 2024 guidance offered multifactor examples, prompting concerns that inventors would need new, AI-specific analysis to justify inventorship.

In contrast, the 2025 document:

• applies the same inventorship test used for all inventions,
• clarifies that AI involvement does not trigger new standards,
• and avoids creating AI-specific burdens for applicants or examiners.

3. It fully revokes the 2024 guidance

The 2024 document remains referenced historically, but no longer governs practice.
FR Doc. 2025-21457 stands as the sole operative standard.

---

Why This Matters for Innovators and Policymakers

The debate over AI-generated and AI-assisted inventions is global—and the U.S. position could influence international discussions at WIPO, WTO, OECD, and domestic legislative agendas.

Photo by Matti Johnson on Unsplash