How to use AI in patent practice: USPTO guidance and compliance tips

As Artificial Intelligence (AI) and large language models (LLMs) increasingly integrate into legal practices, the U.S. Patent and Trademark Office (USPTO) issued guidance to assist patent attorneys with adopting AI tools in patent drafting, prosecution, and other areas of patent law.

In this article, we summarize the key compliance requirements from the USPTO's guidance and explain how Patent Copilot™ helps practitioners meet these obligations while leveraging AI's benefits.

How to use AI in patent practice: USPTO guidance and compliance tips

Core principles

The USPTO has determined that its existing rules sufficiently protect against potential concerns with generative AI use. Importantly:

  • There is no prohibition against using AI in drafting documents, including patent applications, for submission to the USPTO.
  • There is no general obligation to disclose when AI tools have been used.

However, practitioners must remain vigilant about their existing professional obligations when incorporating AI into their workflows.

Signatory, review, and verification requirements

When submitting documents to the USPTO, all submissions must be signed and reviewed by the responsible party, ensuring accuracy and verification of all information. The USPTO emphasizes that simply relying on the accuracy of an AI tool does not constitute a reasonable inquiry under 37 CFR 11.18(b).

How Solve’s Patent Copilot™ helps

Our platform is designed to keep patent professionals in the driver's seat. Patent Copilot™ works collaboratively with attorneys, generating drafts and suggestions that practitioners review and refine; never autonomously filing documents. This workflow ensures that the human attorney maintains full control over verification and review, supporting compliance with signature and certification requirements.

Duty of disclosure requirements

If AI use is material to patentability, it must be disclosed. This includes situations where AI introduces embodiments that inventors did not conceive, or where a named inventor's contributions were actually made by AI. The duty of disclosure cannot be delegated to an AI system. The individuals identified in 37 CFR 1.56(c) bear personal responsibility for ensuring all material information is submitted.

How Solve’s Patent Copilot™ helps

Patent Copilot™ maintains transparency in the drafting process, clearly delineating AI-assisted suggestions from inventor-provided disclosures. This separation helps practitioners assess whether AI contributions rise to the level of materiality and supports proper inventorship analysis throughout the application process.

Confidentiality and client data protection

Under 37 CFR 11.106, practitioners must make reasonable efforts to prevent inadvertent or unauthorized disclosure of client information. General-use AI systems may retain information entered by users and use it to train models. Before using any AI tool, practitioners must understand the tool's terms of use, privacy policies, and cybersecurity practices. Practitioners supervising others must ensure that staff under their supervision also comply with these requirements.

How Solve’s Patent Copilot™ helps

Solve Intelligence™ has built Patent Copilot™ with data security as a foundational principle. We maintain strict confidentiality protocols, and our platform is designed to protect client information. We do not use client data to train models shared with other users, and we provide transparent information about our data handling practices to support practitioners' compliance obligations.

Export control and security concerns

Practitioners must be mindful of foreign filing license requirements and export regulations when using AI tools. AI tools may use servers outside the United States, meaning entered data could be exported in violation of regulations. Foreign filing licenses do not authorize exporting data abroad for preparing US applications. Practitioners must verify their AI tools' infrastructure before use.

How Solve’s Patent Copilot™ helps

Patent Copilot™ is developed with awareness of export control considerations. We provide practitioners with clear, upfront information about our infrastructure and data handling to support their compliance assessments. Our European and US-based teams understand the sensitivity of patent-related technical information and work with practitioners to meet their data security requirements.

Practitioner competence requirements

Under 37 CFR 11.101, practitioners must possess the legal, scientific, and technical knowledge reasonably necessary for representation. The ABA Model Rules, which inform USPTO Rules, specify that lawyers must keep abreast of the benefits and risks associated with relevant technology.

Additionally, 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.

How Solve’s Patent Copilot™ helps

We designed Patent Copilot™ to be intuitive and transparent, helping practitioners understand how AI assists in the drafting process. Our platform includes resources and documentation to support practitioners in developing competence with AI tools, and our collaborative approach facilitates meaningful client communication about AI use in their matters.

Conclusion

The USPTO's guidance provides a comprehensive framework for the responsible use of AI in patent practice. While there is no prohibition on AI use, practitioners must remain attentive to their professional obligations regarding verification, disclosure, confidentiality, export controls, and competence.

How Solve’s Patent Copilot™ helps

At Solve Intelligence™, we are building the first AI-powered platform to assist with every aspect of the patenting process. Our Patent Copilot™ assists with patent drafting, prosecution and office action analysis, portfolio strategy and management, infringement analyses, and litigation workflows.

Patent Copilot™ works with the patent professional; we have designed our products to keep patent professionals in the driving seat. This approach equips legal professionals, law firms, companies, and inventors with powerful tools while maintaining the human oversight that the USPTO's guidance requires.

To learn more about how Patent Copilot™ can enhance your practice while supporting your compliance obligations, contact us today.

AI for patents.

Be 50%+ more productive. Join thousands of legal professionals around the World using Solve’s Patent Copilot™ for drafting, prosecution, invention harvesting, and more.

Related articles

Marbury Law sees 3x-4x efficiency gain from using Solve Intelligence

Key Insights

  • AI adoption requires proof. Bob and his team tested multiple tools before committing, and only moved forward once they saw quantifiable results.
  • 3 to 4x efficiency gains changed the business case. By tracking his own drafting time, Bob demonstrated that AI-enabled workflows made fixed-fee work viable at partner rates.
  • Demonstration drives adoption. Live drafting sessions, client transparency, and side-by-side cost comparisons created full buy-in from both clients and colleagues.
  • Integrated chat removes friction. Keeping research, drafting, and revisions inside one contextual workspace eliminated copy-paste workflows and saved significant time.
  • Context is a force multiplier. AI performs best when it understands the full invention disclosure, file history, and drafting materials in one place.
  • Speed expands strategic value. Faster drafting didn’t just save time - it enabled better coverage, stronger enablement, and real-time responsiveness to client needs.

About Marbury Law

The Marbury Law Group is a premier mid-size, full-service intellectual property and technology law firm in the Washington, D.C. area, with additional strength in commercial law, litigation, and trademark litigation. Recognized by Juristat as a top 35 law firm nationwide and holding Martindale-Hubbell’s AV® Preeminent™ Peer Review Rating, Marbury serves clients ranging from Fortune 500 companies and mid-size technology businesses to high-tech startups and inventors. Its practitioners bring unusually wide-ranging experience, including former technology executives, government R&D managers, startup founders, in-house counsel, “big-law” attorneys, USPTO patent examiners, and judicial clerks. 

Marbury delivers “big-law” service with the flexibility and personal attention of a smaller firm, pairing high-quality work with efficient, budget-aware billing. Based near the USPTO, the firm has drafted and prosecuted thousands of U.S. and foreign patent applications and trademarks, and advises on IP strategy, diligence, and licensing. Formed in 2009 through the merger of two established practices (with roots dating back to 1994), the firm takes its name from Marbury v. Madison (1803), the landmark Supreme Court case that established judicial review.

Introduction

When we sat down with Bob Hansen for this conversation, we knew it would be grounded in both legal depth and real-world business experience. Bob is a founding partner of The Marbury Law Group and has extensive experience across patent prosecution, litigation, licensing, portfolio strategy, and complex IP transactions. But what makes his perspective particularly compelling is that he also brings 20 years of real-world experience as an engineer, program manager, and business executive in Fortune 50 companies and start-ups. He understands firsthand how innovation moves from idea to product, and how intellectual property law fits into that journey.

That dual lens is exactly why we wanted to have this discussion. Bob evaluates technology not just as a patent attorney, but as someone who has managed engineering teams, navigated acquisitions and divestitures, raised capital, and built businesses. When someone with that background says AI has been transformative and backs it up with measurable 3 to 4x efficiency gains, it’s worth listening.

Introducing Solve Review: A Practical Guide to AI-Powered Patent Review

Patent drafting doesn’t end when the first draft is complete. In many ways, the most important work begins at review.

Jurisdictional compliance, internal style alignment, claim clarity, sufficiency of disclosure, and formal requirements. Each aspect of drafting applications must be carefully checked before filing. Yet a thorough review is time-intensive, difficult to standardize, and hard to scale across teams and large portfolios, especially when up against a tight deadline.

Enter Solve Review

With Solve Review, practitioners can run structured, customizable AI-powered reviews in minutes rather than hours, while maintaining transparency, collaboration, and full control over the output. 

Teams using Solve Review report dramatically, with multi-pass manual reviews that previously took three to four hours completing in a fraction of the time

Key benefits

  • AI-powered patent reviews in minutes
  • Each review is fully customizable
  • Save your reviews as templates, run multiple reviews per application
  • Full transparency of working out and results
  • Resolve issues detected by Solve Review with AI

Potter Clarkson Enhances Patent Practice with Solve Intelligence

Solve Intelligence is deployed at Potter Clarkson as a practitioner-led platform, designed to enhance - not replace - the expertise of experienced patent attorneys. The firm uses the technology primarily at a senior level, where skilled practitioners are able to prompt and interrogate the system effectively to guide high-quality outputs.

By combining advanced AI capability with deep technical and legal experience, the platform enables senior attorneys to work more efficiently while focusing their time and judgement on strategic advice, complex analysis and client value. This reflects the firm’s long-standing philosophy that technology should strengthen the role of the practitioner, not substitute professional expertise.

“At Potter Clarkson, our priority is delivering technically rigorous and strategically sound advice to our clients. We use Solve Intelligence as a tool in the hands of experienced patent attorneys - professionals who understand how to guide, challenge and refine AI-generated outputs. It allows our senior teams to concentrate on the aspects of drafting and prosecution where their judgement adds the greatest value, while maintaining full control over quality and client strategy.”

Peter Finnie, Partner, Potter Clarkson

Since rolling out Solve Intelligence’s Patent Copilot, the firm has tailored the platform to reflect its established house styles and drafting standards. This customisation reduces administrative burden and supports consistency across teams, enabling practitioners to engage with AI efficiently without compromising on quality, client-specific requirements, or the firm’s distinctive approach.

Peter Finnie to join Solve's Customer Advisory Board

We are excited to welcome Peter Finnie, Partner at Potter Clarkson, to Solve Intelligence’s Customer Advisory Board.