AI Assistance vs. Inventorship: Ethical Guidelines for Patent Attorneys

What’s the current legal position on AI inventorship?

Patent offices worldwide have consistently rejected AI systems as inventors. The USPTO, European Patent Office (EPO), and UK Intellectual Property Office have all determined that inventors must be natural persons.

The Thaler v. Vidal litigation—in which Dr. Stephen Thaler sought to name his DABUS AI system as an inventor—reinforced this position across multiple jurisdictions. The Federal Circuit held in 2022 that "only a natural person can be an inventor, so AI cannot be."

On November 28, 2025, the USPTO issued revised inventorship guidance that explicitly rescinds its February 2024 framework. The key changes include:

Withdrawal of the Pannu factors framework for AI-assisted inventions.The USPTO acknowledged that Pannu factors "only apply when determining whether multiple natural persons qualify as joint inventors."

Clarification that AI systems are tools, explicitly characterizing them as "instruments used by human inventors," analogous to "laboratory equipment, computer software, research databases, or any other tool that assists in the inventive process."

Affirmation that traditional conception standards apply to all inventions, regardless of whether AI was used in the inventive process.

While this guidance provides welcome clarity, grey areas remain around the boundaries of human contribution when AI plays a significant role in development.

What counts as AI assistance vs. inventorship?

The distinction between AI assistance and inventorship centers on conception—"the touchstone of inventorship" under Federal Circuit precedent. Conception is complete when the inventor has "a specific, settled idea, a particular solution to the problem at hand, not just a general goal or research plan."

AI assistance includes:

• Drafting support for patent applications and claims
• Prior art searches and analysis
• Figure generation and formatting
• Proofreading and consistency checking
• Research acceleration and document review

Inventorship requires:

• Identifying novel concepts and inventive contributions
• Possessing knowledge of all claim limitations
• Having a definite and permanent idea of the complete invention

Consider a practical example: if an attorney uses AI to help draft claim language based on an inventor's disclosed concept, that's assistance. But if an AI system generates a novel solution to a technical problem that a human then simply recognizes as valuable, the human may not have conceived the invention and naming them as inventor could be improper.

The USPTO guidance reaffirms that merely reducing an AI-generated idea to practice does not establish inventorship. The human must contribute to conception itself.

What are the ethical duties of patent attorneys?

The revised guidance does not diminish practitioners' ethical obligations; if anything, it heightens them.

Duty of reasonable inquiry

Under 37 CFR 11.18, every document submitted to the USPTO constitutes a certification that factual contentions have evidentiary support after "an inquiry reasonable under the circumstances." For AI-assisted inventions, practitioners must affirmatively investigate how AI tools were used. Questions to ask clients include:

• What specific problem did the human identify?
• What prompts or inputs did they provide to the AI system?
• How did they evaluate, modify, or build upon AI-generated outputs?
• Did the human possess a definite and permanent idea before or after AI involvement?

Duty of candor and disclosure

Under 37 CFR 1.56, information that "raises a prima facie case of unpatentability due to improper inventorship" must be disclosed. If an attorney learns that a named inventor's contribution was actually made by an AI system, this information may be material and require disclosure even when it adversely affects the client's application.

Competence and confidentiality

Attorneys must understand the tools they use. This includes knowing how AI platforms handle data, whether inputs are used for training, and whether information may be transmitted to servers outside the United States (raising export control concerns). Client confidentiality obligations under 37 CFR 11.106 extend to AI tool selection and use.

Client communication

Under 37 CFR 11.104, practitioners must reasonably consult with clients about the means by which objectives are accomplished, including the use of AI tools in their matters.

Practical guidelines for responsible AI use

To use AI ethically in patent practice, consider these approaches:

• Confirm inventorship with human judgement. Never rely solely on a client's assertion without probing their actual contribution to conception.
• Use AI as an assistant, not an originator. AI should accelerate research and drafting, not replace the inventive process or professional analysis.
• Document everything. Advise clients to contemporaneously record how human inventors conceived claimed inventions, providing evidentiary support for inventorship assertions.
• Maintain transparency with clients. Discuss how AI tools are used and obtain appropriate consent.
• Implement firm-wide policies. Establish guardrails for AI use, particularly for junior attorneys who may not fully appreciate the ethical boundaries.
• Verify data handling practices. Before using any AI tool, understand its terms of use, privacy policies, and server locations.

Common misconceptions clarified

Myth: "If AI contributed, we must name it as an inventor."

False. AI cannot be named as an inventor under current law. The question is whether a natural person conceived the invention, not whether AI assisted in the process.

Myth: "Using AI automatically creates ethical conflicts."

False. The USPTO has confirmed there is no prohibition against using AI in drafting documents and no general obligation to disclose AI tool usage. Ethical use of AI is entirely consistent with professional standards when properly controlled and reviewed by attorneys.

Myth: "The Pannu factors apply to AI-assisted inventions."

No longer accurate. The November 2025 guidance explicitly withdrew this framework, clarifying that Pannu factors only apply when multiple natural persons are involved.

Conclusion

AI will continue to transform patent practice, but inventorship remains a human responsibility. The USPTO's revised guidance confirms that AI systems are tools; valuable instruments that assist human inventors without claiming inventorship themselves.

Attorneys who adopt AI responsibly can gain significant efficiency without crossing ethical or legal boundaries. The key is maintaining human oversight, conducting proper inventorship inquiries, and ensuring compliance with disclosure obligations. Solutions like Solve Intelligence's Patent Copilot^TM exemplify this approach, keeping patent professionals in the driver's seat while accelerating research and drafting workflows.

By treating AI as what it is—a powerful tool rather than a creative originator—practitioners can embrace innovation while upholding the professional standards that clients and the patent system require.