Document Type
Article
Abstract
Over the course of two centuries the US. patent system has generated an expansive compilation of millions of patents, each explaining how to make and use patented technology. Collectively these documents describe the details of inventions in virtually every field of science and engineering, representing a technical database crowdsourced from a global community of inventors. Although this corpus of technical knowledge was originally designed for its role in the U.S. patent system, contemporary Artificial Intelligence ("Al") and Natural Language Processing ("NLP') technologies allow patent text to be used in many productive new ways. In this Article I describe how Al technologies could utilize the patent database to enrich the patent offices, patent owners, and public. AI tools could, for example, use the information contained in millions of patents to answer a wide range of questions about specific technology, provide textual summaries of an area of technology, generate informative abstracts of particular patents, and even assist in drafting patents or technical documents. However, there are legal obstacles to fully realizing this potential. Since many AI-enabled services would require copying and creating derivative works of patent documents, we must consider to what extent copyright might impede these services. Because some AI uses of patent documents are expressive rather than non-expressive, fair use is not as clearly availing as in other AI applications analyzed by legal scholars. I demonstrate that although a patent document constitutes copyrightable subject matter, fair use and other legal doctrines reduce, and might even eliminate, the benefits of copyright protection for patents. Nevertheless, given the chilling effect of the lack of legal clarity, policy changes should be made to remove legal obstacles to the creation of beneficial new AI services.
Recommended Citation
Dean Alderucci,
Copyright Protection for Patents: Some Surprising Implications for Artificial Intelligence,
123
W. Va. L. Rev.
791
(2021).
Available at:
https://researchrepository.wvu.edu/wvlr/vol123/iss3/5