AI Inventorship Post-DABUS: US/EU 2026

AI Inventorship Post-DABUS: US/EU 2026

The headline for AI inventorship patent strategy in 2026 is simple but consequential: in both the United States and Europe, only natural persons qualify as inventors. AI systems, whether foundational models or bespoke engines, remain tools—not inventors. For AI developers, that means mapping each claim element to a human who conceived it, carefully documenting how the system was used, and framing claims around concrete technical improvements that satisfy eligibility standards. Done right, you can secure durable rights while avoiding priority traps and inventorship challenges (https://www.uspto.gov/subscription-center/2025/revised-inventorship-guidance-ai-assisted-inventions; https://www.hklaw.com/en/insights/publications/2026/02/the-human-element-uspto-clarifies-inventorship).

Where US and EU stand on AI inventorship patent rules in 2026

Despite high-profile DABUS test cases, the core rule has not changed: name humans as inventors. In late 2025, the USPTO reverted to traditional conception principles and rescinded its 2024 AI guidance, while the European Patent Office has maintained its 2021 refusal to accept an AI as an inventor. Eligibility analysis for AI-related claims in the US follows the same software law you already know: show the technical improvement under MPEP §2106. In Europe, the DABUS outcomes still control inventorship. None of this is AI-specific inventorship law—AI is treated as a tool (https://www.federalregister.gov/documents/2025/11/28/2025-21457/revised-inventorship-guidance-for-ai-assisted-inventions; https://thompsonpatentlaw.com/artificial-intelligence-patents/).

Why it matters: misnaming an inventor can jeopardize validity and, in the US, claiming priority from a foreign or PCT filing that listed an AI as inventor can raise avoidable headaches unless corrected—so align your naming across jurisdictions from day one (https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).

United States: naming inventors and eligibility post‑DABUS

In the US, the USPTO’s Revised Inventorship Guidance of November 28, 2025 confirms there are no AI-specific inventorship tests. Examiners and applicants must apply traditional conception principles: inventorship turns on who formed a definite and permanent idea of the claimed invention. That means a human who conceived the claimed subject matter is the inventor, even if an AI system assisted; the AI itself cannot be named (https://www.federalregister.gov/documents/2025/11/28/2025-21457/revised-inventorship-guidance-for-ai-assisted-inventions; https://www.uspto.gov/subscription-center/2025/revised-inventorship-guidance-ai-assisted-inventions).

What changed in 2025 was what got rolled back: the USPTO rescinded its February 13, 2024 AI inventorship guidance and returned to uniform conception standards, cautioning applicants about priority risks where earlier or foreign filings listed an AI as inventor. Practically, scrub your priority chain for AI-named filings and correct course before you file in the US (https://ipwatchdog.com/2025/11/26/uspto-issues-new-ai-inventorship-guidance-snubs-vidals-approach/; https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).

On eligibility, AI-related inventions are examined under the same framework as other software claims—MPEP §2106’s two-step analysis. The August 4, 2025 memo to examiners emphasized focusing on concrete technical improvements (for example, reduced latency or memory footprint) rather than aspirational outcomes. Applicants who show the “how” and “why” of improved computer performance fare better (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026; https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).

Signals from leadership also matter. A recent PTAB reversal under Director Squires indicated a more receptive posture toward well‑supported AI eligibility arguments—though the burden remains on applicants to articulate a specific technical improvement tied to claim elements (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026; https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).

Inventorship doctrine on joint inventorship remains human-centered. Federal Circuit precedent limits inventorship to natural persons and requires a contribution to the conception of the claimed subject matter. While that case law developed outside the AI context, its principles carry over: cite the human contributions to conception and treat the AI as an enabling tool—no different than a lab instrument or compiler.

Practical steps for US filings:

• Map each claim to at least one human conceiver. Maintain a contemporaneous record connecting the person’s mental contribution to the claim language.
• Describe the AI’s role as a tool: what inputs were supplied by humans; how parameters, datasets, prompts, or training choices were selected; and how those choices led to the claimed results.
• For §101, draft claims that implement a specific architecture, pipeline, or control mechanism that measurably improves system performance. Explain the mechanism in the specification and use those hooks in prosecution (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026; https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).

Europe after DABUS: what you can and cannot do

The European Patent Office has not shifted since the DABUS matters: only natural persons can be named as inventors, and filings that list an AI as inventor are not accepted. For 2026, applicants should continue naming the human conceivers and proceed under ordinary EPO practice. Align inventorship across your global filings to avoid inconsistencies that can complicate priority or formalities (https://thompsonpatentlaw.com/artificial-intelligence-patents/; https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).

One practical cross-border point: if an early foreign filing named an AI (e.g., to make a point in a test case), plan a clean US and EPO path with human inventors before you claim priority. Do not rely on AI-as-inventor anywhere you plan to pursue protection in 2026.

2026 Compliance Roadmap for AI inventorship patent filings

For in‑house counsel and founders steering AI R&D, the fastest way to de‑risk is to systematize inventorship and eligibility from the start. Use this roadmap:

1) Identify human conceivers per claim

• Before drafting, list each claim element and the human(s) who conceived it. Keep dated notes, design docs, or tickets linking the idea to a person.
• Capture why the person’s mental step mattered—e.g., a new training regimen, a novel control loop, or a data‑path optimization (https://www.uspto.gov/subscription-center/2025/revised-inventorship-guidance-ai-assisted-inventions).

2) Treat AI as a tool—and document it

• Record prompts, hyperparameters, and dataset choices; note who chose them and why.
• Store versioned artifacts: model cards, evaluation reports, and ablation studies showing the human’s intentional design choices (https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).

3) Frame claims on concrete technical improvements

• Articulate measurable benefits like reduced inference latency, lower memory use, improved throughput, or better robustness—and tie them to claimed mechanisms (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026; https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).

4) Scrub priority across jurisdictions

• Avoid naming AI in any filing that could sit in your priority chain. If a foreign or PCT filing did so, consult counsel to plan corrections or new filings before US/EPO prosecution to protect priority (https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).

5) Prosecute under standard rules—and cite current guidance

• In the US, argue eligibility using MPEP §2106 and the August 2025 memo’s focus on technical improvements. In replies, reference the 2025 Revised Inventorship Guidance as needed (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026; https://www.federalregister.gov/documents/2025/11/28/2025-21457/revised-inventorship-guidance-for-ai-assisted-inventions).

6) Monitor fast‑moving eligibility developments

• Track PTAB and Federal Circuit outcomes touching AI; the Director’s recent reversal signals a more practical approach to well‑supported AI claims (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026).

Drafting strong AI claims: show the “how,” not just the result

Most eligibility misfires in AI patent practice stem from abstract, result‑oriented claiming. You can avoid them with a few drafting habits:

• Specify the architecture: name the layers, modules, or control elements and how they interact (e.g., a scheduler that gates token generation based on a learned latency model) rather than claiming “optimized inference.”
• Tie the claim to a concrete improvement: “achieving sub‑50ms median latency on edge hardware by pruning and quantization coordinated via a feedback controller,” not “faster model.”
• Disclose the mechanism in the spec: include ablations and benchmarks that isolate the claimed technique as the cause of the improvement.
• Anchor improvements to computer functionality: reduced cache misses, fewer memory swaps, less network chatter—effects examiners recognize as technical (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026).

A helpful pattern:

• Independent claim: a system/process implementing your core control or data‑path innovation.
• Dependent claims: concrete variants—parameter ranges, cache layouts, quantization steps, scheduler policies, or data‑fusion rules.
• Method and computer‑readable medium claims: mirror the system with steps and storage, to cover multiple infringement modes.

Back it up with evidence. While not required, including data in the spec that shows the improvement—“X% reduction in latency on Y hardware from Z mechanism”—gives the examiner a clear hook to allow the case. Even if the numbers shift later, the mechanism‑first story remains intact (https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).

Common pitfalls, fees, and planning for AI portfolios in 2026

Avoid these recurring issues we see across AI teams:

• Naming an AI (e.g., DABUS) as an inventor anywhere in your priority chain. This creates avoidable formalities issues and can undermine US strategy (https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).
• Applying rescinded 2024 “significant contribution” tests or mixing them with joint inventorship factors not tailored to AI‑assisted work. The 2025 guidance returned to traditional conception standards—use those (https://ipwatchdog.com/2025/11/26/uspto-issues-new-ai-inventorship-guidance-snubs-vidals-approach/; https://www.uspto.gov/subscription-center/2025/revised-inventorship-guidance-ai-assisted-inventions).
• Drafting abstract, result‑oriented claims that never show how the improvement is achieved. Examiners will push back unless you articulate the technical mechanism (https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026).
• Failing to document human conception and the AI’s tool‑like role. If you cannot connect a person to each claim, your inventorship story is vulnerable (https://www.hklaw.com/en/insights/publications/2026/02/the-human-element-uspto-clarifies-inventorship).
• Overlooking cross‑border filing risks that complicate US priority. Clean up naming before you file in the US or EPO (https://www.patentnext.com/2026/03/the-usptos-revised-guidance-regarding-ai-assisted-inventions/).

Fees and timing in 2026:

• No AI‑specific fee schedule exists. Standard USPTO fees apply. For example, a small‑entity utility application’s filing fee is roughly in the ~$320 range, with separate search and examination fees also due. Budget for the full stack of government fees and professional costs (https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).
• Expect rising activity: AI‑related filings have surged in recent years, with notable growth across technology subclasses and a significant share of generative‑AI families reaching publication by 2023—reinforcing the value of filing early and often where you have defensible improvements (https://thompsonpatentlaw.com/artificial-intelligence-patents/; https://www.gtlaw.com/en/insights/2026/1/ai-patent-outlook-for-2026).

Portfolio planning tips:

• Align inventorship and eligibility reviews with sprint demos. Treat every major architecture or pipeline change as a potential disclosure event.
• File provisionals to capture fast iterations; consolidate into non‑provisionals once data stabilizes.
• Coordinate with privacy and data‑licensing leads so your training‑data story is ready for diligence.

The bottom line on AI inventorship patent practice in 2026

Across the US and EU, AI systems are still not inventors. For every AI inventorship patent filing in 2026, your job is to: name the human(s) who conceived each claim element; document the AI as a tool; and draft claims that concretely improve computer performance, explained in the specification. Stay aligned with the USPTO’s November 2025 guidance, the examiner focus on technical improvements under MPEP §2106, and the EPO’s continued post‑DABUS stance. When you operationalize these habits, you minimize priority risk and maximize allowance odds (https://www.federalregister.gov/documents/2025/11/28/2025-21457/revised-inventorship-guidance-for-ai-assisted-inventions; https://www.venable.com/insights/publications/2025/12/the-101-reset-for-2026; https://thompsonpatentlaw.com/artificial-intelligence-patents/).

Get Help From GTC

Building with AI and need a clear, defensible patent story? Global Trademark Company helps AI teams map human inventors to claims, draft eligibility‑proof specifications, and coordinate US/EPO filings that avoid priority traps. Visit /services/patent to schedule a consult, or email hello@globaltrademarkcompany.com. Let’s turn your next AI breakthrough into a durable patent—without tripping over inventorship rules.