Introduction
The evolving landscape of AI-assisted inventions presents unique challenges and opportunities in the realm of patent law. Recent guidance from the United States Patent and Trademark Office (USPTO) addresses the complexities of patenting AI and software technologies, particularly in the context of co-inventorship involving AI systems. This guidance aims to clarify the criteria for patent eligibility and provide examples to assist legal practitioners in navigating this intricate field.
Description
A natural person can qualify as a co-inventor alongside an AI system [3], making the invention patentable with the human inventor recognized as the sole inventor [3]. The United States Patent and Trademark Office (USPTO) has issued updated guidance on the subject-matter eligibility of AI-assisted inventions [2], effective February 13, 2024 [4]. This guidance clarifies the criteria for determining whether an invention constitutes an abstract idea and how it can be integrated into a practical application [2], addressing the challenges of patenting software and AI technologies since the Digital Revolution.
Key aspects of the guidance include examples that illustrate improvements to technological processes through AI [2], such as automating previously unautomatable tasks [2], enhancing error correction [2], improving data flow monitoring, and advancing medical device analysis [2]. The USPTO has introduced specific examples of patent-eligible AI subject matter [2], including the use of artificial neural networks for anomaly detection [2], AI methods for analyzing speech signals [2], and AI models for personalizing medical treatments based on individual patient characteristics [2]. These examples aim to assist legal practitioners in navigating patent applications related to software and AI innovations [2], aligning with President Biden’s Executive Order 14110 [4].
In the life sciences sector, the integration of AI and machine learning (ML) presents significant opportunities for advancements in diagnostics [5], therapeutics [5], and personalized medicine [5]. A recent study involving 1,018 scientists at a large US firm specializing in materials science highlights the role of AI in selecting and advancing candidate compounds for various applications [1], including healthcare and industrial manufacturing [1]. Scientists utilized simulations and domain knowledge to predict material characteristics [1], rank candidate compounds based on their predicted quality [1], and synthesize the most promising options to evaluate their true properties [1]. However, securing patent protection for AI/ML-based inventions in this field remains challenging [5], particularly due to frequent rejections by the USPTO under 35 USC § 101, which addresses patentable subject matter [5]. Inventions at the intersection of AI/ML and life sciences often face heightened scrutiny [5], with patent examiners rejecting claims on the grounds of being directed to abstract ideas or natural phenomena [5]. The Supreme Court’s decision in Mayo Collaborative Services v [5]. Prometheus Laboratories (2012) underscores the difficulty of patenting natural laws [5], such as drug dosage correlations [5], unless they are accompanied by additional inventive steps [5].
The development of a specialized AI model requires a natural person to design [3], develop [3] [5], and train the AI to address a specific problem [3], focusing on generating outputs with desired properties [3]. Identifying these properties is crucial as it links the problem to the AI’s solution approach [3]. Proper documentation of all steps taken to specialize the AI model is essential to demonstrate the contributions made by the scientists [3]. The process involves evaluating [3], refining [1] [3], and testing candidate compounds generated by the AI [3]. Refinement implies that scientists modify the AI-generated compounds [3], ensuring that the final outputs align with the intended application [3]. This highlights the importance of human oversight and expertise in the AI development process [3], particularly in the context of patentability and intellectual property rights [3].
To enhance the chances of securing patent protection [5], life sciences companies should collaborate with experienced patent counsel to craft claims that emphasize the tangible benefits and transformative applications of AI/ML in biological and clinical settings [5]. Additionally, the alignment of the USPTO’s inventorship guidance with AI-assisted innovation is crucial for those employing AI in the innovation process who seek patent protection for their results [1]. A webinar on the new guidance will be held on March 5, 2024 [4], where USPTO personnel will provide an overview and address stakeholder questions [4], further emphasizing the evolving landscape of AI-assisted inventions and the necessity of developing robust patent strategies to maintain a competitive advantage in this rapidly evolving industry.
Conclusion
The USPTO’s updated guidance on AI-assisted inventions marks a significant step in addressing the complexities of patenting in the digital age. By providing clear criteria and illustrative examples, the guidance aims to facilitate the patenting process for AI and software technologies. This development has profound implications for industries such as life sciences, where AI and ML are driving innovation. Companies are encouraged to work closely with patent experts to navigate the challenges and leverage the opportunities presented by AI-assisted inventions, ensuring they remain competitive in a rapidly evolving technological landscape.
References
[1] https://www.lexology.com/library/detail.aspx?g=a44642b4-26e2-4cf6-894d-16ccc20c550e
[2] https://www.lexology.com/library/detail.aspx?g=cffdff5e-0321-4167-ad8d-5ad140ee1a9a
[3] https://www.jdsupra.com/legalnews/mit-study-aligns-with-uspto-s-3537630/
[4] https://www.uspto.gov/about-us/events/inventorship-guidance-ai-assisted-inventions-webinar
[5] https://natlawreview.com/article/overcoming-patent-challenges-aiml-assisted-life-sciences-techbio-inventions
