Introduction
This is the first installment in a two-part series examining the implications of using generative artificial intelligence (GAI) in the drafting and prosecution of patent applications. In this part, we address privilege and discovery risks that could arise when GAI tools are used in the patent-drafting process, and we identify targeted discovery strategies that patent litigators should consider when challenging patents that may have been drafted with GAI assistance. Later, in part two, we turn to the potential impact of having drafted a patent application with a GAI tool, including potential challenges of invalidity for anticipation and written description. And, in part two, we offer practical guidance for patent prosecutors and inventors seeking to use GAI in their workflows.
The Ubiquity of GenAI in Patent Workflows
As GAI platforms become increasingly integrated into research and development workflows – from brainstorming new inventions, to conducting prior-art searches, to even drafting patent specifications and claims – both patent prosecutors and litigators are evaluating novel privilege and discovery issues that have already started sprouting in other areas of the law.
Courts are beginning to grapple with the intersection of GAI and legal privilege, and the law in this area is still evolving. Some courts have suggested that communications with GAI platforms may not enjoy the protections of the attorney-client privilege or the work-product doctrine. Under this view, because the attorney-client privilege traditionally protects confidential communications between a client and their lawyer made for the purpose of obtaining legal advice, and GAI is not a lawyer, interactions with GAI may fall outside these protections. However, other courts may take a different approach, potentially analogizing GAI tools to other technological aids that attorneys have long used without waiving privilege. This question remains unsettled and will likely be refined as courts continue to address it.
Depending on how courts resolve these questions, there may be consequences in patent prosecution that could bear out later in litigation. Namely, because the GAI platform is neither a lawyer nor litigation-aware, feeding privileged information into a GAI platform may constitute a voluntary disclosure to a third party – thereby waiving the privilege altogether. So, amidst a backdrop of uncertainty as to how courts will evaluate the use of GAI tools in the patent-drafting process, patent prosecutors should carefully consider the potential implications of integrating GAI tools into their workflows.
Privilege and Discoverability Considerations for AI-Generated Materials
GAI tools are becoming an integral part of the research and development process for many businesses. Inventors are beginning to use GAI platforms to flesh out their ideas, help research prior art, craft patent specifications, and even draw figures for patent applications. Whether and to what extent GAI interactions qualify for privilege protection remains an open question that courts have not definitively resolved.
When patent prosecutors use GAI platforms in the drafting process, questions arise about the discoverability of these materials. When a patent prosecutor inputs technical disclosures, invention records, or strategic assessments into a GAI platform, some may argue that input is being shared with a third-party service provider. Depending on the GAI platform used, an opposing party might contend that such disclosure affects the confidentiality necessary to maintain privilege. However, the outcome of such arguments may depend on factors including the specific platform used, applicable data-processing agreements, and how courts ultimately address these novel issues.
The work-product doctrine presents additional considerations. Work product must be prepared "in anticipation of litigation," a standard that patent-prosecution materials – drafted in anticipation of obtaining a patent, not litigating one – will often fail to satisfy. Even where a prosecutor might argue that certain prosecution strategies were informed by potential future enforcement, the doctrine's protections are unlikely to extend to the GAI platform's outputs themselves, which are generated by a third-party tool rather than reflecting the mental impressions of an attorney.
The potential consequence is that prompts submitted to a GAI tool, the GenAI tool's responsive outputs, and any iterative exchanges between the user and the GAI tool may be subject to discovery requests in subsequent litigation. Whether such materials will ultimately be producible may depend on factual circumstances, the specific GAI platform and its terms of service, and how courts continue to develop the law in this area.
Discovery Considerations in Patent Litigation
In light of these evolving considerations, both parties in patent litigation should be aware of potential GAI-related discovery issues. For patent litigators defending against infringement claims or challenging a patent's validity, the following discovery considerations may be relevant, though the ultimate discoverability of such materials will depend on the specific facts and applicable privilege determinations:
- Requests Related to Patent Drafting. Discovery requests may seek all prompts, inputs, and outputs associated with any GAI tool used in drafting or revising the patent specification, claims, or drawings. This could include technical information, invention disclosures, or inventor communications that were provided to a GAI platform as part of the drafting process, as well as any GAI-generated drafts, figures, or portions thereof that were incorporated into the application as filed. The discoverability of such materials may depend on the privilege and work-product determinations discussed above.
- Requests Related to Office Action Responses. During prosecution, practitioners may use GAI to help formulate arguments distinguishing prior art, to draft claim amendments, or to develop legal and technical reasoning set forth in office action responses. Discovery requests may seek GAI-generated materials relating to the prosecution of the patent-in-suit, including materials used to formulate responses to rejections under Sections 101, 102, 103, and 112 of the Patent Act. As with other materials, the discoverability of such information would depend on applicable privilege determinations.
- Requests Related to Subject-Matter Eligibility. Responses to Section 101 rejections may receive attention in discovery. When a patent examiner rejects claims under 35 U.S.C. § 101, typically on the ground that the claims are directed to an abstract idea, law of nature, or natural phenomenon without an inventive concept, the applicant's response often involves articulating why the claims recite something more than the abstract idea. If a GAI tool were used to help formulate those arguments, the resulting prompts and outputs may be subject to discovery requests.
- Requests Related to Prior-Art Searches. GAI-powered prior-art search tools are becoming increasingly common. Discovery requests may seek the search parameters, prompts, and results generated by any GAI tool used to conduct prior-art searches during prosecution. Whether the use of GAI-powered search tools affects this analysis is an open question, and practitioners on both sides should be aware of these considerations.
- Requests Targeting Internal Policies and Communications. Beyond the GAI interactions themselves, discovery requests may seek internal policies, guidelines, or training materials governing the use of GAI tools in patent prosecution, as well as communications among attorneys, patent agents, and clients regarding the use of GAI tools in connection with the patent-in-suit. The discoverability and relevance of such materials would depend on the circumstances of the case.
- Interrogatories and Deposition Topics. In addition to document requests, interrogatories may ask the patentee and its counsel to identify each GAI tool used during prosecution, the specific tasks for which each tool was used, the individuals who interacted with the tools, and the extent to which GAI outputs were incorporated into the filed application or prosecution submissions. These topics may also be relevant for Rule 30(b)(6) depositions.
Conversely, patent litigators asserting infringement claims against an accused infringer should be aware that similar GAI-related discovery considerations may apply to accused infringers. The following discovery considerations may be relevant:
- Requests Related to Design-Around Efforts. If the accused infringer used a GAI platform to analyze the patent claims and generate alternative designs or workarounds, the discoverability of such materials would depend on the applicable privilege determinations.
- Requests Related to Willfulness and State of Mind. If the accused infringer used a GAI tool to evaluate infringement risk or to develop defenses, those materials may be relevant to questions of subjective knowledge and willfulness, subject to applicable privilege protections.
- Requests Related to Development of the Accused Instrumentality. Where a GAI platform was used to develop, refine, or optimize the accused technology, the content of those interactions may be relevant to various issues in the case. The discoverability and relevance of such materials would depend on the circumstances.
- Requests Related to the Accused Infringer's Internal Policies and Communications. As with discovery directed to the patent holder, internal policies, guidelines, or communications governing the accused infringer's use of GAI tools in product development or patent clearance may be relevant. Again, the discoverability and relevance of such materials would depend on the circumstances of the case.
Given the potential for discovery requests related to the use of GAI, both patent prosecutors and litigators should give advance consideration to these issues. Patent prosecutors may wish to discuss with their clients the potential implications of using GAI in research and development efforts, and litigators should ensure that litigation hold notices adequately encompass prompts, inputs, and outputs relating to the use of GAI. Part Two of this series will elaborate on practical considerations surrounding the use of GAI in patent prosecution and litigation.
Conclusion
In sum, the use of GAI tools by both patent prosecutors and inventors alike raises questions that may arise in future litigation. If it is learned, through discovery, that a GAI tool was used in authoring any part of the patent application, parties may dispute whether such materials are protected by the attorney-client privilege or work-product doctrine. These questions are likely to be addressed by courts in the coming years as GAI use becomes more prevalent in patent practice.
For questions regarding the use of generative artificial intelligence, intellectual property strategy, or risk mitigation, please contact Nicole Berkowitz Riccio, Dominic Rota, or any member of Baker Donelson's Intellectual Property Group.