Introduction
The complexities of inventorship in the realm of AI-assisted inventions present unique challenges, particularly concerning the determination of significant contributions by inventors. This issue is further complicated by the evolving nature of AI technology and its role in the inventive process, raising questions about intellectual property rights and patent validity.
Description
An inventor must make a significant contribution to the conception or reduction to practice of an invention [1], which cannot be deemed “insignificant in quality” relative to the invention’s full scope [1]. This requirement is particularly complex in the context of AI-assisted inventions [1], where the determination of inventorship can lead to disputes [1]. The USPTO emphasizes the duty to disclose any evidence of improper inventorship [1], and patent examiners may inquire about inventorship even if it does not affect patentability [1]. This is especially pertinent for life sciences companies collaborating with external parties [1], as IP ownership often hinges on US inventorship laws [1].
Recent discussions among key stakeholders, including legal advisors and representatives from major corporations and academic institutions [2], have highlighted the current state of AI technology in the invention creation process [2]. These discussions focus on how to handle inventions significantly contributed to by AI [2], as well as the unanticipated intellectual property challenges arising from AI-driven innovation [2]. Legal experts are particularly concerned with the implications of black-box AI on the inventive process and whether applicants need to disclose AI contributions in patent applications to maintain patent validity [2].
The USPTO guidance attributes significant contributions to individuals based on their roles in developing and training AI systems [1]. Ownership of an AI system does not dictate inventorship of its outputs [1]. Companies must evaluate the type of AI system utilized [1], distinguishing between off-the-shelf systems for general problems and bespoke systems tailored for specific issues [1]. Training a bespoke AI system can be considered a significant inventive contribution [1], while general training of an off-the-shelf system typically does not meet this threshold [1].
Human interaction in training AI systems can constitute a significant contribution [1], particularly when developing prompts to elicit specific outputs [1]. This practice [1], known as “prompt engineering,” is recognized by the USPTO as potentially significant [1], although guidance on what constitutes an inventive prompt is limited [1]. Simply reformulating a general problem into an AI prompt does not qualify as a significant contribution; more sophisticated prompts that target specific responses are necessary [1].
Documentation of prompt engineering efforts is essential [1], especially when collaborating with third parties [1]. Companies should establish clear guidelines and identify individuals responsible for prompt engineering [1]. When using off-the-shelf AI systems [1], claiming inventorship based solely on AI-generated outputs is insufficient [1]. Human modifications to improve AI-generated compounds can be significant [1], but claims should be carefully tailored to specific [1], validated compounds rather than broad categories [1].
Characterization of AI-generated outputs in statements of work is crucial [1]. If further validation is required to select specific compounds from a list of AI-generated candidates [1], this should be clearly indicated [1]. As AI continues to influence drug discovery [1], challenges regarding inventorship and ownership are expected in pharmaceutical litigation and prosecution [1]. Thorough documentation of human contributions [1], including the development and training of AI models and subsequent modifications [1], will be vital in addressing these challenges [1].
Conclusion
The integration of AI in the inventive process necessitates a reevaluation of traditional inventorship criteria, particularly in the context of patent law. As AI technology continues to advance, the implications for intellectual property rights and the legal frameworks governing them will become increasingly significant. Companies must adapt to these changes by ensuring meticulous documentation and clear guidelines to navigate the complexities of AI-assisted inventorship and maintain the integrity of their intellectual property claims.
References
[1] https://www.jdsupra.com/legalnews/inventorship-in-the-age-of-ai-6836655/
[2] https://www.uspto.gov/about-us/events/aiet-partnership-series-3-ai-driven-innovation
