On April 11, 2024, the United States Patent and Trademark Office (USPTO) issued Guidance on the Use of Artificial Intelligence-Based Tools (Guidance). This applies existing rules and policies to the use of artificial intelligence (AI) in practice. . This guidance was issued based on President Biden's October 2023 Executive Order on the Safe, Secure, and Trustworthy Development and Use of Artificial Intelligence, which provides government agencies with protections against AI-related harms. We are requesting that the law be enacted and enforced. Multiple government agencies are currently issuing guidance based on the Executive Order. The USPTO's new guidance does not introduce new regulations, but instead discusses the risks associated with using AI to create documents, file documents, and interact with USPTO systems. We offer suggestions for mitigating those risks. Address confidentiality and national security implications. Additionally, the USPTO's duty of good faith and good faith to applicants and practitioners, including the duty to disclose known information material in order to obtain patentability, covers actions performed using AI tools. It reminds me of that.
Draft documents to be submitted to the USPTO
This guidance provides specific suggestions for reviewing AI-generated content for inclusion in USPTO filings. For example, the guidance states that when preparing briefs or responses to office actions, practitioners should check the accuracy of citations and ensure that all arguments are legally and technically sound. states. When using AI to draft patent applications, practitioners should verify technical accuracy, compliance with 35 USC 112(a), and appropriate distinction between prospective and working examples in the application text. must pay close attention to. Additionally, when using AI to develop disclosure statements, practitioners should verify that all cited technologies are relevant and that no relevant technologies are omitted.
Practitioners should also avoid submitting AI-generated images or specimens as evidence of trademark use when filing new applications, maintaining and renewing trademark registrations, and responding to office actions. yeah. Additionally, a practitioner should be careful when he submits AI-generated evidence in U.S. Trademark Trial and Appeal Board proceedings. AI-generated images are considered misrepresentations of fact because they do not represent the actual use of the trademark in the marketplace. Furthermore, AI-generated materials submitted as evidence that contain irrelevant information or are unnecessarily accumulated may also be discarded as inappropriate.
Use AI to interact with the USPTO's electronic systems
AI systems and tools are not considered “users” for purposes of submitting or accessing documents through the USPTO's electronic filing system. Therefore, the AI cannot obtain his USPTO.gov account, have a USPTO Patent Center account, be sponsored as a support staff member, or sign applications. Additionally, AI systems should not be used in a manner that generates mass access to USPTO databases or violates the USPTO's Terms of Use.
Confidentiality and national security concerns
According to the guidance, practitioners should ensure that AI providers' terms of service, privacy policies, and cybersecurity practices protect sensitive information from inadvertent disclosure or export. To do so, practitioners need to ensure that their AI systems adequately protect their data, do not use client information to train models, and do not send client information to third parties. there is. Practitioners will also need to obtain permission before allowing AI tools to send data overseas. Please note that the rights granted by the Foreign Filing License are limited to international filings and do not provide the necessary export licenses for AI tools.
Duty of candor and honesty
This guidance provides numerous examples of how the USPTO's duties of honesty and good faith apply to the use of AI. For example, practitioners must ensure that their use of AI does not violate their obligations to have the legal, scientific, and technical knowledge to represent their clients. A practitioner must also conduct their own reasonable investigation when signing her correspondence with the USPTO, as relying on AI tools does not amount to reasonable investigation.
Although there is no general requirement to notify the USPTO of AI uses, the USPTO requires patents to disclose significant AI uses. When using AI to draft claims for a patent application, a practitioner must notify her USPTO if the claimed subject matter lacks significant contributions by human inventors. This also occurs when an AI system helps draft a detailed description of a patent application and introduces alternative embodiments that ultimately fall within the scope of the patent claims but lack significant human contribution. There is likely to be. The implication of a claim that is significantly lacking in human contribution is that “patent protection is [only] Inventions in which a natural person made a significant contribution are required. ” Inventorship Guidance for AI-Assisted Inventions, 89 FRB. Registration 10043 (February 13, 2024); See also Saylor vs. Vidal43 F.4th 1207, 1213 (Fed. Cir. 2022) (“[O]AI cannot be an inventor because only natural persons can be inventors. ”)1 Patent applicants who plan to use AI to support technology development must not only comply with disclosure obligations to the USPTO but also protect themselves from future invalidation attempts by others regarding potential failures. You should consider internal documentation processes to enable rebuttals. Revealing human involvement, or insufficient human involvement.
The U.S. Patent and Trademark Office's position on human inventorship requirements is consistent with the U.S. Copyright Office's position on human authorship, and the Office has issued similar guidance regarding the application of AI-generated materials and claims. Requires sufficient human authorship to support it. Copyright Registration Guidance: Works Containing Material Generated by Artificial Intelligence, 88 Fed. Registration 16190 (March 16, 2023); See also Saylor vs. PerlmutterCase No. 22-1564 (DDC, August 18, 2023) (“Human authorship is a fundamental requirement of copyright.'').
