Navigating the Use of AI in Patent Practice: USPTO Issues Comprehensive Guidance for Practitioners


 Navigating the Use of AI in Patent Practice: USPTO Issues Comprehensive Guidance for Practitioners

Have you ever utilized AI to search for an invention or related prior art? What about employing generative systems like Microsoft’s ChatGPT or Google’s Gemini to draft portions of a patent application or arguments to overcome rejections from the United States Patent and Trademark Office (USPTO)? With the emergence and widespread adoption of AI, the USPTO has intensified its efforts to address AI-related considerations in innovation and intellectual property. On April 11, 2024, the USPTO released a comprehensive guidance document titled “Guidance on Use of Artificial Intelligence-Based Tools in Practice Before the United States Patent and Trademark Office” (89 Fed. Reg. 25609). Aimed at patent lawyers, agents, and other practitioners utilizing AI tools, the guidance serves as a roadmap, emphasizing existing rules and offering strategies to mitigate associated risks.

Understanding Existing Rules and Policies:

The guidance begins by reiterating practitioners’ obligations under existing USPTO rules and policies. Key among these are the duty of candor and good faith (37 C.F.R. §§ 1.56, 1.555, 42.11), which applies across all interactions with the USPTO, not just patent examination contexts. Practitioners are also reminded of their duty of confidentiality and the requirements associated with the USPTO’s policy on foreign filing licenses and electronic systems. The guidance provides specific examples to illustrate how these rules apply in practice.

Duty to Disclose Information Material to Patentability:

One significant aspect addressed in the guidance is the “duty to disclose all information—including on the use of AI tools by inventors, parties, and practitioners—that is material to patentability.” While there is no blanket obligation to disclose the use of AI, practitioners must disclose when its involvement rises to the level of materiality under 37 C.F.R. § 1.56(b). For example, disclosure is required if an inventor does not make a “significant contribution” to any one of the claims submitted in a patent application, but instead relies upon an AI to generate the subject matter for that claim. Additional guidance regarding inventorship for AI-assisted inventions may be found in the USPTO’s February guidance—Inventorship Guidance for AI-Assisted Inventions, 89 Fed. Reg. 10043 (Feb. 13, 2024).

Navigating the Signature Requirement:

The guidance also emphasizes the importance of the signature requirement and corresponding certifications in ensuring document integrity. Documents drafted with AI assistance must be reviewed and personally signed by the practitioner, certifying their accuracy and authenticity. This is crucial to prevent potential abuses, as current AI capabilities can “hallucinate” or “confabulate” information, leading to the fabrication of false authority. One illustration the guidance includes is the use of AI to populate an information disclosure statement (IDS), which may result in large numbers of cumulative and irrelevant prior art submissions. The guidance cautions practitioners that such IDS submissions containing irrelevant and marginally pertinent cumulative information could be construed as papers presented for an improper purpose, running afoul of 37 CFR 11.18. Thus, practitioners must avoid relying solely on AI-generated content and conduct a reasonable inquiry to ensure the accuracy of citations and legal arguments.

Confidentiality Concerns:

Regarding confidentiality, practitioners are urged to exercise diligence in preventing the disclosure of sensitive information while utilizing AI systems for tasks such as conducting prior art searches or drafting patent applications. It is crucial to acknowledge that AI systems possess the capability to retain data input by users, potentially leveraging it for model refinement or sharing it with external entities, thereby contravening practitioners’ obligations to maintain confidentiality. When enlisting third-party services for AI development or data storage, practitioners bear the responsibility of ensuring the preservation of confidentiality.

Practical Guidelines for AI Use:

The guidance also provides practical guidance for using AI systems to file documents with the USPTO. Practitioners must ensure that AI tools do not exceed their authorized access and comply with USPTO filing systems’ requirements. For example, AI systems are not considered “users” under USPTO filing systems and cannot be used to file follow-on documents in pending applications. Moreover, practitioners must exercise caution to avoid unauthorized access or data mining of USPTO systems, adhering to terms of use and avoiding potential violations.


In summary, the USPTO’s guidance offers a comprehensive framework for navigating the use of AI in patent practice. By emphasizing practitioners’ obligations, disclosure requirements, and strategies for mitigating risks associated with AI utilization, the guidance promotes transparency, integrity, and responsible AI use in the patent system. Practitioners who incorporate these principles into their practice will not only be better equipped to comply with the USPTO’s policies and procedures but will also be well-positioned to effectively utilize AI systems and enhance efficiency for their clients.

Samuel Raque | Associate Attorney