The “Genesis” Gap: How the New Executive Order Collides with Patent Law’s Human Requirement
Will AI Agent-led Research be Patentable?
The “Genesis Mission” Executive Order (EO) does more than just allocate funding; it effectively mandates a method of scientific discovery that sits in a legal gray zone. By explicitly directing the Department of Energy (DOE) to deploy “AI agents” for “autonomous... experimentation” (EO Sec. 3(a)(vi)), the White House is pushing federal research beyond the boundaries of current patent eligibility.
For IP owners and patent practitioners, this creates an immediate crisis of strategy. The EO presumes the existence of “innovations arising from AI-directed experiments” (EO Sec. 5(c)(ii)), yet under current U.S. law and USPTO guidance, such innovations may trigger ownership claims by multiple parties—or potentially fall into the public domain immediately upon creation.
The Conflict: “AI-Directed” vs. “Human Conception”
The core of the conflict lies in the definition of the creative act. The USPTO’s Inventorship Guidance for AI-Assisted Inventions establishes that while AI-assisted inventions are not categorically unpatentable, the “inventorship analysis should focus on human contributions, as patents function to incentivize and reward human ingenuity” (89 FR 10043, Summary).
The Guidance relies heavily on the Pannu factors, reiterating that “inventorship is limited to a natural person(s)” (89 FR 10043, Sec. II). The USPTO is explicit that “conception is the touchstone of inventorship” and that because conception is a mental act, “it has to date been understood as only performed by natural persons” (89 FR 10043, Sec. III.B).
Contrast this with the “Genesis Mission” directives:
EO Sec. 1 commands the creation of AI agents to “test new hypotheses” and “automate research workflows.”
EO Sec. 3(a)(ii) defines these agents as tools to “explore design spaces” and “evaluate experimental outcomes.”
EO Sec. 3(a)(vi) points to “autonomous and AI-augmented experimentation.”
If an AI agent on the American Science and Security Platform (Platform) independently formulates a hypothesis, explores the design space, and evaluates the result to identify a new superconductor, where is the human inventor?
Under the USPTO’s Guiding Principle 2 (89 FR 10043, Sec. IV.B), the user of the Platform likely fails to qualify as an inventor, as “merely recognizing a problem... does not rise to the level of conception.”
DOE Messaging: Framing AI as “Instrument”
While the Executive Order leans heavily into “autonomous” terminology, the Department of Energy’s November 24, 2025, press release frames the initiative differently—likely a strategic choice that softens the potential legal-syntactical conflict.
In the release, “Energy Department Launches ‘Genesis Mission’ to Transform American Science and Innovation Through the AI Computing Revolution,” Secretary of Energy Chris Wright describes the Genesis Mission as “giving our scientists and engineers tools to work at the speed of innovation.”
Similarly, Under Secretary for Science Dr. Darío Gil characterizes the Platform not as an independent agent, but as “a scientific instrument for the ages.”
This subtle distinction—”agent” vs. “instrument”—may be critical for patent practitioners.
The “Agent” Narrative (EO): Implies the AI does the work independently. This creates a “Thaler problem” where the invention has no human conceiver.
The “Instrument” Narrative (DOE): Implies the AI is merely a sophisticated tool, akin to a microscope or telescope. If the Platform is a “scientific instrument,” then the human scientist using it retains inventorship, likely satisfying the USPTO’s requirement that a human “contribute in some significant manner to the conception” (89 FR 10043, Sec. IV.A).
By positioning the Genesis Mission as a productivity multiplier (“double the productivity,” according to the press release) rather than a replacement for human discovery, the DOE appears to be aligning its messaging with the current “human-centric” patent framework.
However, inventors and counsel should be wary: if the function of the technology (autonomous experimentation) contradicts the description (tool), the USPTO and courts will likely look to the function.
The Reduction to Practice Trap
Even if the DOE’s “instrument” narrative holds, the technical reality of the Platform poses risks. The EO envisions AI agents that “evaluate experimental outcomes” (EO Sec. 3(a)(ii)).
However, the USPTO warns in Guiding Principle 3 regarding reduction to practice:
“Reducing an invention to practice alone is not a significant contribution that rises to the level of inventorship... Therefore, a natural person who merely recognizes and appreciates the output of an AI system as an invention... is not necessarily an inventor” (89 FR 10043, Sec. IV.B).
Consequently, a scientist who uses the Genesis Platform to find a molecule, where the AI did the “exploring” and “evaluating,” may find herself legally unable to sign an inventor’s oath, as she did not contribute to the conception.
The Ownership Quadrilemma
EO Section 5(c)(ii) directs the Secretary of Energy to “establish clear policies for ownership... including innovations arising from AI-directed experiments.” However, without a legislative update to the Patent Act, agency policy cannot override the Federal Circuit’s interpretation in Thaler v. Vidal.
This creates a complex four-way ownership dispute for any breakthrough achieved on the Platform:
The User: Claims ownership based on their prompt. But the USPTO states that “maintaining ‘intellectual domination’ over an AI system does not, on its own, make a person an inventor” (89 FR 10043, Guiding Principle 5).
The Developer: The creator of the “foundation models” (EO Sec. 3(a)(iv)) may claim the output. The USPTO notes in Guiding Principle 4 that the person who “designs, builds, or trains an AI system... could be an inventor” (89 FR 10043, Sec. IV.B), though this is fact-dependent.
The Federal Government: If the invention lacks a human inventor, the government may assert that as a product of federal resources (e.g., “synthetic data generated through DOE computing resources” (Sec. 3(a)(v))), the innovation is effectively public property or subject to “march-in rights.”
The Public Domain: If the invention lacks a human who made a “significant contribution” to the conception, the USPTO must reject the application under 35 U.S.C. 101 and 115.
Regulatory and Trade Secret Implications
Because patent protection may be unavailable for these “AI-directed” discoveries, the EO suggests a pivot toward “trade-secret protections” (Sec. 5(c)(ii)). This is a logical fallback but presents practical hurdles.
Trade secrets require secrecy. Yet, the Platform is a “shared access” environment (Sec. 6(a)(ii)) involving “student researchers” (Sec. 6(a)(iii)) and “external partners” (Sec. 5(c)). Maintaining the requisite secrecy to enforce trade secret rights on a government-run, multi-tenant cloud platform will be logically difficult.
If the “provenance tracking” (Sec. 3(d)(i)) mandated by the EO inadvertently discloses the method of discovery to a cleared federal contractor, the trade secret status could be compromised.
Analysis & Thoughts
Benefits:
Forcing the Issue: The EO’s directive for “clear policies” (EO Sec. 5(c)(ii)) forces a confrontation between the executive branch’s scientific goals and the judicial branch’s patent interpretation.
Contractual Clarity: The requirement for “standardized partnership frameworks” (EO Sec. 5(c)(i)) suggests a move toward contract-based IP rights rather than reliance solely on patents.
Challenges and Risks:
Unpatentable Subject Matter: IP counsel must warn clients that using AI tools like the Platform for “autonomous” discovery may result in unpatentable assets. The USPTO guidance is clear: “Applications and patents must not list any entity that is not a natural person as an inventor... even if an AI system may have been instrumental” (89 FR 10043, Sec. III).
Duty of Disclosure: Practitioners have a “duty to disclose to the USPTO information that raises a prima facie case of unpatentability due to improper inventorship” (89 FR 10043, Sec. V.B(i)). Using the Genesis Platform may trigger this duty, requiring applicants to disclose that the “conception” was performed by a federal AI agent.
Global Disadvantage: If other jurisdictions (e.g., the UK or China) update their laws to clearly recognize AI-generated inventions while the US adheres to the “natural person” requirement, the Genesis Mission could ironically drive IP ownership offshore.
Conclusion
The Genesis Mission Executive Order assumes that AI should be the primary driver of scientific discovery. The current USPTO guidance assumes it cannot be the inventor.
Until this disconnect is resolved—likely through the “clear policies” promised in Section 5 or future legislation—IP counsel should advise owners to view Platform-generated research as highly vulnerable.
As the USPTO emphasized, “patents function to incentivize and reward human ingenuity” (89 FR 10043, Summary). The Genesis Mission, by design, seeks to incentivize machine efficiency—a goal that currently has minimal safe harbor in the Patent Act.
Inventors and practitioners will be wise to watch how the Department of Energy implements the Genesis Mission.
Disclaimer: This is provided for informational purposes only and does not constitute legal or financial advice. To the extent there are any opinions in this article, they are the author’s alone and do not represent the beliefs of his firm or clients. The strategies expressed are purely speculation based on publicly available information. The information expressed is subject to change at any time and should be checked for completeness, accuracy and current applicability. For advice, consult a suitably licensed attorney and/or patent professional.