Introduction
Artificial intelligence has become an indispensable tool in contemporary research and development. This technological transformation raises a foundational legal question: under what circumstances does the use of AI still result in a human invention, and when does it cross the threshold into subject matter that lacks human inventorship and is therefore not patentable?
Patent systems traditionally rest on the premise that exclusive rights reward human intellectual achievement. As AI systems increasingly participate in — and sometimes autonomously generate — technical solutions, this premise is tested.
U.S. and European patent law address this challenge in markedly different ways. U.S. law treats human inventorship as a substantive legal requirement whose absence may affect validity and enforceability. German and European patent law, by contrast, maintain a formal natural-person requirement but do not attach validity consequences to the absence of an actual human inventive contribution once a patent has been granted.
The following sections analyze these divergent approaches.
I. The U.S. Approach: Human Inventorship as a Substantive Requirement
1. Statutory Framework and Case Law
Under 35 U.S.C. § 100(f), an inventor must be an “individual,” i.e., a natural person. This interpretation was confirmed by the Federal Circuit in Thaler v. Vidal, 43 F.4th 1207 (Fed. Cir. 2022), which held that AI systems cannot qualify as inventors under U.S. patent law.
On November 28, 2025, the USPTO published its Revised Inventorship Guidance for AI-Assisted Inventions (Federal Register Doc. 2025-21457), rescinding its February 13, 2024 guidance in full. The revised guidance confirms that no separate inventorship test applies to AI-assisted inventions; traditional inventorship principles govern all applications.
Only natural persons may be named as inventors.
2. Inventorship as a Substantive Condition of Validity
In U.S. law, inventorship is not merely a procedural designation. Naming an individual who did not contribute to the conception of the claimed invention may constitute:
• inequitable conduct,
• and potentially lead to invalidity or unenforceability.
In litigation, an accused infringer may argue that the named inventor did not actually conceive the invention and that the claimed subject matter lacks qualifying human contribution. The issue is therefore doctrinal and evidentiary.
3. Qualifying Human Contributions in AI-Assisted Contexts
U.S. practice acknowledges that meaningful human inventorship may exist even where AI systems play a substantial technical role. What matters is whether the human contribution rises to the level of conception. Depending on the circumstances, the following activities may carry inventive significance:
- Identifying the technical problem, defining constraints, and setting the direction of development may constitute conception if these activities reflect inventive insight.
- In advanced AI systems, the formulation of prompts, the selection of model parameters, training constraints, or evaluation criteria may embody technical judgment and creative choice.
- Where an AI system produces multiple potential solutions, the human selection of a particular solution based on technical reasoning may amount to an inventive contribution.
- Human adaptation, combination, or refinement of AI-generated outputs may establish inventorship where such intervention involves inventive decision-making. The decisive point is evidentiary: these contributions must be demonstrable. Without documentation, inventorship may not be defensible.
4. Practical Consequences: Documentation as Defensive Infrastructure
For applicants seeking U.S. protection, documentation of human inventive contribution is strategically essential. Recommended measures include:
• contemporaneous technical records,
• documentation of prompt design and output evaluation,
• records of human reasoning and decision-making,
• formal inventorship review prior to filing.
U.S. law thus embeds human inventorship as a substantive precondition for enforceable patent rights.
II. The European and German Approach: Formal Requirement Without Validity Consequences
1. Natural-Person Requirement
Under § 6 PatG, Article 81 EPC, and Rule 19 EPC, only natural persons may be designated as inventors. The German Federal Patent Court confirmed this in its DABUS decision (BPatG, 11 November 2021, 11 W (pat) 5/21). AI systems cannot be named as inventors under German or European patent law.
2. Inventorship as a Personal Right
Despite this formal requirement, inventorship is treated as a personal right rather than a substantive patentability criterion. Established doctrine provides:
• Incorrect inventorship is not a ground for opposition (Art. 100 EPC; § 59 PatG).
• Incorrect inventorship is not a ground for nullity (Art. 138 EPC; § 22 PatG).
• The nullity grounds are exhaustive.
• Violation of the right to be named as inventor may justify correction or reassignment, but not invalidity.
The same structure applies under Article 65 UPCA in Unified Patent Court (UPC) proceedings.
3. Consequences in AI Contexts
As a result, a patent may remain valid and enforceable even where the designated inventor did not in fact make an inventive contribution. This creates structural tensions:
• Patent exclusivity may depend primarily on access to AI systems rather than human inventive activity.
• The framework may discourage transparent disclosure of AI involvement.
• Existing remedies presuppose the existence of a true human inventor and provide no solution where no human inventive act exists.
As AI technologies continue to evolve, the divergence between formal inventorship rules and substantive validity is poised to acquire tangible economic significance. It may generate a competitive imbalance between the U.S. and European patent regimes, since applicants seeking protection for AI-generated inventions may effectively benefit from a more permissive validity framework in Europe.
4. Reform Considerations
Possible reform directions include:
• Introducing lack of human inventive contribution as a ground for opposition and nullity;
• Moving toward a substantive concept of inventorship;
• Establishing evidentiary standards for AI-assisted innovation.
At a minimum, the current framework requires substantive reconsideration in cases where an invention is demonstrably generated without human inventive activity, as the existing doctrinal structure no longer adequately reflects technological reality.
Conclusion
AI systems are increasingly capable of generating technical solutions with minimal human intervention. Patent law must respond in a manner that preserves its foundational premise: the reward of human intellectual achievement. The United States treats human inventorship as a substantive requirement affecting validity and enforceability. German and European patent law maintain a formal natural-person requirement without attaching validity consequences.
Whether this divergence remains sustainable is an open and pressing question. A structured, evidence-based concept of human inventive contribution may ultimately be necessary to safeguard the normative foundations of the patent system.
Talk to us and contact us for further information at:
Phone: +49 (0)611 / 341568-0
E-mail: Send email to us
PGP-Key
Author: Markus Richardt