Innovations in Patents: AI's Role from Drafting to Examination – Insights from the WIPO and Future Considerations


Artificial intelligence (AI) is a general-purpose technology that is significantly influencing today’s digital economy and society. In recent years, AI has exhibited remarkable and incomparable advancements in technology, resulting in a fundamental transformation of the landscape of innovation and intellectual property (IP). The exponential growth of AI technology poses numerous complex challenges to the established principles of patent law. As such, AI presents a set of complex legal quandaries that have not yet been thoroughly examined in the context of patent law, an area that is significantly impacted by these technological breakthroughs. As a result, there may be far-reaching consequences for IP, and talks are already underway about how emerging technologies will alter the IP   landscape. Therefore, it is essential to strike a delicate equilibrium between promoting technological innovation and preserving legal integrity. This write up explores the intricate relationship between AI advancements and legal integrity within the realm of patent law which is indispensable considering the current context. It delves into the ways AI has revolutionized the innovation process, influencing how patents are created, examined, and enforced. Furthermore, it sheds a light on the use of AI tools in patent prior art searching, automated patent drafting & filing. In short, it examines the problems, prospects, and potential impact of generative AI technology on the IP system.

AI for patent prior art searching

The determination of patentability is the essential factor in the awarding of patent rights to a novel innovation. The process of granting a patent consists of multiple stages that assess whether an innovation meets the criteria for being eligible for a patent.[1] An innovation is deemed unpatentable if it does not satisfy the necessary criteria. [2]  Patent prior art searching is an initial and highly challenging task, akin to finding a needle in a haystack. The primary objective of conducting a patent prior art search is to identify the most pertinent prior art documents from a vast pool of published patent applications that already exists. In essence, it seeks out previous examples of innovation that could affect the eligibility of a patent application. [3]  Figure 1 illustrates the process of conducting a prior art search.

Figure 1 Steps in Prior Art Search

The illustration given (figure-1) depicts multiple stages of conducting a prior art search. As indicated in the diagram, steps 2 and 3 are the most significant. Developing the search statement necessitates a thorough comprehension of the crucial subject matter and the probable originality of the application.[4] In many instances, examiners revise the search statement iteratively based on their comprehension of the previous art or the possible patentability of the application.[5] 

Much like the process of formulating a search statement, it is of utmost importance to carefully examine the numerous patents that have been obtained in order to extract relevant information.[6] 

It is, perhaps, the most time-consuming stage. Therefore, taking into account the required level of technical expertise, patent-prior art searching is regarded as the most suitable method for enhancing the time efficacy of this task involving the retrieval of information. Conversely, algorithmic intelligence is employed in the post-search analysis to scrutinize drawers, sticky notes, and color coding/highlighting. The main linguistic[7]  and semantic[8] obstacles here include legal wordings, complex sentences, acronyms, and the technical character of claims.[9] Therefore, it is indisputable that the most immediate and conspicuous effects of in the field of patent law are the improvements made to prior art searches, in which vast databases of scientific literature, patent documents, and other sources are efficiently analyzed by AI algorithms, streamlining the process and diminishing the effort and time needed to conduct a thorough prior art search. As the capacity of human beings to analyze immense amounts of data is limited, this technology not only decreases examination time but also enhances its quality. The United States Patent and Trademark Office (USPTO) has recently incorporated a new function called "Similarity Search" into the Patents End-to-End (PE2E) search suite.[10] This feature, based on AI, aims to aid patent examiners in conducting prior art searches.

Automated Patent Drafting & Filing

Recent years have seen an increase in the number of patent attorneys employing AI tools to automate the drafting and evaluation of patent applications. By proposing pertinent prior art, identifying potential challenges, and optimizing claim language, these tools unequivocally aid innovators and patent attorneys in generating more comprehensive and sturdy patent applications. These innovations not only enhance the efficacy of patent applications but also augment the likelihood of obtaining successful patent grants. Furthermore, it has greatly enabled patent attorneys and agents to automate the drafting and prosecution of patents. Inventors have the option to utilize automation technologies to facilitate the drafting of their patent application, either by a patent attorney or a patent agent. Several instances of AI are presently employed for automated drafting and prosecution. Starting with the LegalTech[11] startups have created technology solutions to automate the process of drafting patent applications.[12]  Another example is, Specifio, a platform that offers automated patent drafting software for inventions. It can quickly convert a practitioner's collection of method patent claims into an initial draft of a patent application in only a few minutes.[13] TurboPatent is a cloud-based platform that uses a proprietary AI-powered drawing tool to automate the process of producing and prosecuting patent applications.[14] It simplifies the process of responding to office actions including arguments made in response to rejections from patent examiners. Thus, this tool, reduces the time required by a paralegal from half an hour to just a few minutes.[15] In the United States, TurboPatent is being used for presenting the USPTO's office action in a user-friendly format. Other than these AI, ANAQUA StudioTM[16] is also used in drafting. It offers patent drafting tools that effectively cut drafting time in half, assess patent drafts for compliance with statutory requirements, and present found flaws. The use of such tools, including software tools, by inventors is not prohibited by US patent law.[17]

Again, AI automation is currently being employed for defensive patenting purposes.[18] For instance, Cloem is a startup that employs its artificial-intelligence technology to generate numerous Cloems, which are computer-generated variations of probable alternative definitions, based on a supplied claim.[19] Therefore, it is clear that AI technologies can significantly help practitioners with the preparation of response arguments and, in certain instances, with the provision of automated responses in technical domains where patent examiners frequently reject claims pertaining to business-methods patent claims.[20]

The WIPO’s Stance & future consideration

As lawmakers have started to understand the far-reaching effects of AI and how it can disrupt the current IP system, WIPO has started to zero in on AI's IP-specific features.  Notably, WIPO Translate and WIPO Brand Picture Search are two AI applications that utilize AI-based algorithms for automated translation and picture identification.

In May 2018, WIPO convened a summit to discuss the use of AI and to promote the dissemination and collaboration about these applications.[21] In this summit, the purpose of WIPO was to create a comprehensive list of the most urgent difficulties and concerns regarding the protection of IP that have emerged due to the increasing usage of AI as a versatile technology. After a year, a meeting was held at WIPO in September 2019 with the aim of facilitating a dialogue among member States and representatives from the commercial, research, and non-governmental sectors.[22] It deems that the increasing proliferation of AI technologies necessitates novel approaches within the existing patent law system. In this context, only a novel procedure and a new framework for safeguarding the possible outcomes of AI, are capable of tackling the challenges that posed by AI that is functionally equivalent to humans (referred to as 'strong AI').


While AI holds promise as a highly effective tool for conducting prior art searches, analysis, and drafting, its results are not exempt from constraints. In certain instances, particular legal issues demand the application of specialized assessments or detailed, thorough analyses. In such scenarios, solely relying on complete automation may overlook the genuine essence of the subject matter.  To avoid these challenges, AI can be employed to identify relevant information that impacts a legal decision, while it does not exclusively determine it. However, prior to doing the analysis, it is crucial to assess the average individual's expertise in the relevant field, including their grasp of specific terms and concepts.[23]


[1] Chimuka Garikai, ‘Impact of artificial intelligence on patent law. Towards a new analytical framework- the Multi-Level Model’, [2019] 59 World Patent Information 101926

[2] Ibid.

[3] Setchi Rossitza, Irena Spasic et al., ‘Artificial Intelligence for Patent Prior Art Searching’ [2021] 64 Word Patent Information 102021

[4] S. Bashir, A. Rauber, Improving retrievability of patents in prior art search, in: C. Gurrin, et al. (Eds.), Advances in Information Retrieval. ECIR 2010. Lecture Notes in Computer Science, vol. 5993, Springer, Berlin, Heidelberg, 2010.

[5] F. Crestani, Combination of similarity measures for effective spoken document retrieval, Journal of Information and Science. 29 (2) (2003) 87–96; S. Adams, Is the full text the full answer? – considerations of database quality, World Patent Inf. 54 (2018) S66–S77.

[6] C.D. Manning, P. Raghavan, H. Schütze, Introduction to Information Retrieval, Cambridge University Press, 2008.

[7] C. Fellbaum, WordNet: an Electronic Lexical Database, MIT Press, Cambridge, MA, 1998.

[8] I. Spasic, S. Ananiadou, J. McNaught, A. Kumar, Text mining and ontologies in biomedicine: making sense of raw text, Briefings Bio inf. 6 (3) (2005) 239–251.

[9] A. Krishna, B. Feldman, J. Wolf, G. Gabel, S. Beliveau, T. Beach, Examiner Assisted Automated Patents Search. AAAI Fall Symposium Series: Cognitive Assistance in Government and Public Sector Applications, 2016.

[10] USPTO, New PE2E Search Tool Using AI Search Features (Jan 11, 2023), <> accessed Feb 2, 2024.

[11] Sarah Garber, The Third Wave: Why Big Data is the Future of Legal Tech, IPWATCHDOG (Aug. 9, 2016), < > accessed 19 January, 2024.

[12] Catalyst Investors, LegalTech is Primed for Growth Investments, ROSS BLOG (Dec. 1, 2017), <> accessed 10 January, 2024; Eva Hibnick, What is Legal Tech?, L. INSIDER BLOG (Sept. 7, 2014) (copy on file with Georgia State University Law Review)

[13] SPECIFIO, [] (Accessed 20 January, 2024)

[14] Taylor Soper, This Startup Just Launched Software to Automate Patent Application Process, GEEKWIRE (Apr. 23, 2015), <> accessed 16 January 2024; TURBOPATENT, <> accessed 22 January, 2024.

[15] Robert Ambrogi, New AI-Powered Patent Tool Helps Prepare Responses to Office Actions, LAWSITES (Feb. 22, 2017), <> accessed 15 November, 2023.

[16] ANAQUA, < >Accessed 11 November, 2023.

[17] U.S. Patent. No. 8,041,739, at [1] (filed Aug. 31, 2001).

[18] Bill Barrett, Defensive Use of Publications in an Intellectual Property Strategy, 20 NATURE BIOTECHNOLOGY, Feb. 2002, at 191, 191,

[19] Technology, CLOEM, <> accessed  Feb.1, 2024.

[20] ibid



[23] Kinetic Concepts, Inc. v. Blue Sky Med. Grp., Inc., 554 F.3d 1010, 1022 (Fed. Cir. 2009) (accepting evidence regarding how a doctor of ordinary skill would understand the claim).

Jannatul Shareat Disha, is currently pursuing her Ph.D. in International Law at Zhongnan University of Economics and Law in Wuhan, China. She is a recipient of the prestigious China Scholarship Council (CSC) fully funded scholarship, a testament to her academic excellence and dedication to advancing her expertise in the field. Disha’s journey in law began at East West University, where she earned her LLB degree, laying the foundation for her stellar academic career. She further honed her legal acumen by obtaining an LLM from Bangladesh University of Professionals. Notably, she has been recognized as a Belt and Road Scholar, a distinction earned through her commitment to the study and application of international legal principles. Adding to her impressive academic repertoire, Disha pursued an MBA at Beijing Normal University, specializing in Belt and Road Initiative, where she was again granted a fully funded scholarship by the Chinese Government. Prior to embarking on her Ph.D. journey, Disha contributed significantly to the legal landscape during her three-year tenure in corporate practice at MCLaw Services. 

Post a Comment