USPTO Inventorship Guidance Reshapes Carrier AI Patent Filing Strategy

From tracking USPTO filings by the top 20 insurance carriers since 2022, the inventorship guidance change is already visible in filing patterns: applications that previously required extensive human-contribution documentation now proceed with standard inventor declarations. The November 28, 2025 revision, published in the Federal Register as Document No. 2025-21457, formally rescinded the February 2024 guidance issued under former Director Kathi Vidal and replaced it with a unified framework that treats AI systems the same as any other tool in the inventive process.

The practical significance for insurance companies is substantial. Under the prior guidance, a data scientist at State Farm who used a machine learning system to discover an optimal claims-triage architecture had to demonstrate, claim by claim, that their contribution met the heightened "significant contribution" standard borrowed from joint inventorship law. Under the new guidance, that same inventor simply needs to show they conceived the invention in the traditional sense: that they possessed a definite and permanent idea of the complete invention. The AI system, whether it generated candidate architectures, optimized parameters, or identified patterns in training data, is classified identically to a microscope, a calculator, or a database query tool.

What the Old Guidance Required

The February 13, 2024 guidance issued under Director Vidal applied the Pannu v. Iolab Corp. factors (155 F.3d 1344, 1351, Fed. Cir. 1998) to AI-assisted inventions. The Pannu test, originally designed to evaluate whether multiple human contributors qualify as joint inventors, imposed three requirements: (1) significant contribution to the conception or reduction to practice of at least one claim, (2) contribution of non-insignificant quality measured against the full scope of the invention, and (3) doing more than merely explaining well-known concepts or the current state of the art.

Applying this framework to single-inventor AI scenarios created an unusual analytical burden. Patent applicants had to distinguish their human contributions from the AI system's outputs as though the AI were a potential joint inventor whose share needed to be separated. For insurance carriers filing applications describing AI-assisted underwriting models, pricing algorithms, or claims automation systems, this meant extensive documentation showing exactly which elements the human inventor conceived independently of the AI's output.

The practical effect was threefold. First, patent prosecution costs increased because applicants needed detailed invention disclosure forms documenting the human contribution at each stage. Second, examiners had a new basis for rejection: failure to demonstrate adequate human contribution relative to the AI's role. Third, strategic ambiguity around what constituted "significant contribution" when working with generative AI systems made risk assessment for patent portfolios more difficult.

For carriers like USAA, which has filed patents describing generative AI systems for clarifying aerial imagery in property damage assessment, the old standard created particular friction. When a generative model produces enhanced imagery that reveals previously undetectable damage patterns, documenting exactly where the human inventor's conception ends and the AI's novel output begins required careful legal work that added months and thousands of dollars to each filing.

The New Unified Standard: AI Purely as a Tool

The November 2025 guidance, issued under Director John A. Squires, eliminates the Pannu overlay entirely for single-inventor cases. The traditional conception test now governs all inventions uniformly, regardless of whether AI systems assisted in the inventive process. Conception is defined as "the formation in the mind of the inventor, of a definite and permanent idea of the complete and operative invention, as it is hereafter to be applied in practice."

The guidance explicitly classifies AI systems as "analogous to laboratory equipment, computer software, research databases, or any other tool that assists in the inventive process." This language deliberately places generative AI, large language models, and machine learning systems on the same legal footing as a spreadsheet or a wind tunnel. No separate or modified legal standard exists for AI-assisted inventions.

The key inquiry is whether the natural person "possessed knowledge of all of the limitations of the claimed invention such that it is so clearly defined in the inventor's mind that only ordinary skill would be necessary to reduce the invention to practice, without extensive research or experimentation." When one natural person creates an invention with AI assistance, the analysis stops there. When multiple natural persons are involved, traditional joint inventorship principles (including the Pannu factors) apply only among the human contributors.

The guidance implements Executive Order 14179 (January 23, 2025), "Removing Barriers to American Leadership in Artificial Intelligence," which directs federal agencies to review and revise prior administration policies that impede U.S. AI leadership. It applies uniformly across utility, design, and plant patents.

Director Squires and the Pro-Innovation Philosophy

John A. Squires, sworn in as the 60th USPTO Director on September 22, 2025, has articulated a substantially different vision for AI patent prosecution than his predecessor. At the AIPLA Annual Meeting on November 14, 2025, Squires stated: "I firmly and without reservation believe in a strong, robust, expansive, and resilient intellectual property system." He added that "categorically excluding AI innovations from patent protection jeopardizes America's leadership in this critical emerging technology."

The philosophical shift extends beyond inventorship to examination practice. On August 4, 2025, Deputy Commissioner for Patents Charles Kim issued a memo to patent examiners in software-related technology groups (including AI and ML). The memo warned against overusing "mental process" rejections for machine learning claims, emphasized that algorithms processing volumes of data "cannot practically be performed" in human minds, and instructed examiners to consult specifications showing how inventions improve technology. Kim set the standard that rejections should only occur when it is "more likely than not" that a claim is ineligible.

On September 26, 2025, the Appeal Review Panel (Squires, Acting Commissioner Valencia Martin Wallace, and Vice Chief Administrative Patent Judge Michael W. Kim) decided Ex parte Desjardins, vacating a PTAB decision that had rejected machine learning claims as patent ineligible. The panel confirmed that claims showing "improvement in computer, or improvement to other technology" are patent eligible. On December 5, 2025, Kim announced an advance notice of MPEP revision incorporating the Desjardins holding into examiner guidance.

Squires summarized the aggregate effect at AIPLA: "The doors to America's Innovation Agency are wide open. We are open for business." For insurance carriers evaluating whether to expand their patent filing programs, this represents the clearest signal since the Alice decision in 2014 that the USPTO intends to be receptive to AI-related applications.

Which Carrier Patent Portfolios Benefit Most

According to Insurance Journal's December 2025 analysis of Evident (AI benchmarking platform for financial services) data, three P&C insurers dominate the insurance AI patent landscape: State Farm with 326 AI patents since 2014, USAA with 218, and Allstate with 136. Combined, these firms hold 77% of all insurance industry AI patents. P&C insurers collectively account for 89% of all insurer AI patents filed since 2014.

The concentration matters for analyzing the inventorship guidance impact. Carriers with the largest existing portfolios face the most pending applications and continuation filings that previously required Pannu-compliant documentation. State Farm's 326-patent portfolio spans machine learning-driven claims triage, autonomous vehicle fault analysis, and predictive pricing models. Each new application in these families previously needed to clear the heightened inventorship standard; now they proceed under the simpler traditional conception test.

The data shows broader industry trends as well. Among 30 major insurers across North America and Europe, 166 AI patents have been filed since January 2023. Generative AI patents surged from 4% to 31% of filings between 2014 and October 2025. Only three insurers have filed agentic AI patents, with USAA leading. Patent activity peaked in 2020 and has not fully rebounded, creating a potential filing window for carriers that accelerate under the new guidance.

Alexandra Mousavizadeh, CEO of Evident, noted that "patents offer a rare window into where insurers are placing their biggest bets on AI. This data shows that innovation is overwhelmingly being driven by a handful of U.S. firms, especially in P&C." The inventorship guidance change disproportionately benefits these same U.S. firms because the prior Pannu-based standard was a uniquely American legal construct with no equivalent in other patent jurisdictions.

Carrier	AI Patents (Since 2014)	Key Filing Areas	Inventorship Guidance Impact
State Farm	326	Claims triage, autonomous vehicle fault, pricing ML	Largest portfolio; most pending continuations affected
USAA	218	Generative AI imagery, agentic AI, property damage	GenAI filings benefit most from simplified standard
Allstate	136	In-vehicle AI, behavior-based pricing, claims automation	Active filing program in telematics AI models
All other insurers	~200 (combined)	Various	Lower volume reduces absolute benefit

The Critical Distinction from Section 101 Eligibility

The inventorship guidance addresses who can be named as an inventor. Section 101 patent eligibility addresses whether the invention itself constitutes patentable subject matter. These are separate legal hurdles, and the November 2025 guidance only resolves the first.

The Alice/Mayo two-step analysis for patent eligibility remains unchanged in federal courts. Under this framework, examiners first determine whether a claim is directed to an abstract idea, law of nature, or natural phenomenon. If so, they evaluate whether the claim contains an "inventive concept" sufficient to transform the abstract idea into patentable subject matter. For insurance AI patents describing underwriting algorithms, pricing models, or risk classification systems, this remains the primary prosecution obstacle.

The Federal Circuit's April 2025 decision in Recentive Analytics v. Fox Corp. illustrates the continuing challenge. The court invalidated AI patents that applied generic machine learning to entertainment scheduling, holding that "merely implementing an abstract idea using generic machine-learning models on conventional computing infrastructure does not represent technological improvement sufficient to move claims out of the realm of abstract ideas." The court emphasized that "merely applying established machine learning methods to a different field" does not create patent-eligible inventions.

Venable LLP's December 2025 analysis ("The Section 101 Reset for 2026") notes that despite USPTO receptiveness under Squires, "patents granted under current examination may still face invalidity challenges in litigation under existing court doctrine." This creates an asymmetry: carriers can more easily obtain AI patents under the current examination environment, but those patents may prove vulnerable if challenged in court.

The practical two-hurdle framework for insurance carriers now looks like this: the inventorship hurdle (can we name our data scientist as inventor when they used AI tools?) has been simplified. The eligibility hurdle (is our AI-assisted claims-triage system patent-eligible subject matter rather than an abstract idea?) remains complex. The Squires-era USPTO has signaled favorable treatment on both fronts, but the Federal Circuit has not followed on the eligibility question.

Strategic Implications for Carriers Building AI Patent Portfolios

The simplified inventorship standard creates several tactical opportunities for insurance companies actively filing AI patents.

Reduced Documentation Burden

Under the old guidance, carriers needed detailed invention disclosure forms documenting human contributions at each stage of AI-assisted development. KTS Law's December 2025 analysis recommends that applicants still "document human conception by recording inventive thought, the specific problem, hypothesis, and core solution as encoded in AI prompts," but this documentation now serves as a best practice for potential future disputes rather than a prosecution requirement. The shift from mandatory to optional reduces filing costs and timeline for each application.

Provisional Application Strategy

The timing creates a filing window advantage. Patent activity across the insurance industry peaked in 2020 and has not fully recovered. Carriers that file provisional applications now, while the simplified standard is in effect and before competitors respond, can lock in priority dates for AI innovations developed during the 2023-2025 generative AI acceleration period. A provisional application requires only a written description; claims and the formal inventorship analysis can be deferred for twelve months.

Continuation and Divisional Filings

Existing patent families benefit immediately. State Farm's 326-patent portfolio likely has dozens of pending continuation applications where the inventorship analysis was conducted under the old Pannu-based standard. These applications can now proceed with simplified inventor declarations. For carriers with large pending portfolios, the guidance eliminates a class of potential office actions that would have required responses (at $3,000-$8,000 per response) explaining the human inventor's contribution relative to the AI system.

Generative AI Filing Acceleration

The shift from 4% to 31% generative AI patents between 2014 and October 2025 signals where the industry is heading. Generative AI inventions, where a human directs an AI system to produce novel outputs, were the most problematic under the old standard because the line between human conception and AI generation is often blurred. The new guidance eliminates this analytical difficulty entirely: if the human inventor conceived the complete invention (directed the AI, selected from its outputs, integrated elements into a working system), traditional inventorship is satisfied regardless of how much creative work the AI system performed.

Documentation Best Practices Under the New Standard

Despite the simplified prosecution path, Brownstein Hyatt's December 2025 analysis warns that "companies face heightened USPTO scrutiny and potential litigation challenges" and that "robust evidence of human conception will be demanded in disputes." Mayer Brown similarly recommends demonstrating how human contributors: (1) selected and specified AI-identified features, (2) integrated those elements into claimed solutions, and (3) possessed requisite understanding of all limitations.

For insurance carriers, this means maintaining lab notebooks or digital equivalents showing the actuary or data scientist's problem formulation, hypothesis, solution architecture, and inventive judgment applied in selecting or modifying AI outputs. This documentation is not required for prosecution but protects the patent in inter partes review or district court litigation.

The Gap Between Examination and Litigation Standards

Greenberg Traurig's March 2026 follow-up analysis, "Recalibrating AI Patent Strategy: What the USPTO Shift Means and What It Doesn't," identifies a critical risk: "Companies treating these changes as a general invitation to file broadly, without precision, may find themselves with a portfolio that looks strong until it is tested."

The gap manifests in two ways. First, the USPTO's examination standard (reject only when "more likely than not" that a claim is ineligible) is more applicant-friendly than the federal court standard for invalidity (clear and convincing evidence). A patent that survives examination may not survive litigation. Second, the inventorship simplification removes a prosecution barrier but does not strengthen the underlying patent against invalidity challenges based on inadequate human contribution.

For insurance carriers investing millions in patent prosecution, the strategic question is whether the current USPTO environment should drive accelerated filing or whether the eventual litigation standard should constrain strategy. Greenberg Traurig concludes that successful portfolio building "requires experienced IP counsel who understand both the technical substance of AI innovations and prevailing court doctrine, not simply taking advantage of lenient examination standards."

From tracking the intersection of patent law and insurance innovation since the first AIG AI patent filings in 2016, the prudent approach for carriers is to file aggressively (taking advantage of the simplified inventorship standard and favorable examination posture) while drafting with litigation in mind (including technical specificity, concrete improvements to model architecture, and clear documentation of human conception).

Why This Matters for Actuarial Practice

The inventorship guidance change intersects with actuarial work in several direct ways. First, many of the AI patents held by State Farm, USAA, and Allstate describe systems that actuaries helped develop: predictive pricing models, loss ratio optimization algorithms, claims-severity classifiers, and telematics-based risk segmentation tools. The actuaries and data scientists who directed AI systems to develop these tools are the inventors whose status the guidance clarifies.

Second, the competitive dynamics of patented pricing technology affect actuarial consulting and rate filing. When a carrier patents a novel ML-based pricing approach, competitors cannot replicate it without license or design-around. The concentration of 77% of insurer AI patents at three carriers suggests an emerging IP moat around certain actuarial technologies that could widen as filing accelerates under the new standard.

Third, for actuaries working in model validation and governance, patented AI systems create documentation requirements that overlap with ASOP No. 56 (Modeling) obligations. A patented model's specification, claims, and prosecution history become part of the governance record. The inventorship standard simplification means fewer required documents in patent prosecution, but actuarial model governance still demands comprehensive documentation of model development, testing, and validation regardless of the patent filing standard.

The bottom line: the three carriers that dominate insurer AI patents now face a lower barrier to expanding those portfolios. For competing carriers, actuarial consultancies, and insurtech vendors developing similar technologies, the window to establish defensive patent positions has narrowed. The combination of simplified inventorship, favorable examination posture under Squires, and the Subject Matter Eligibility Declarations (available since December 4, 2025) creates the most patent-friendly environment for insurance AI innovations since before Alice in 2014.