IP lifecycle management using blockchain and machine learning: Application to 3D printing datafiles
Graphical abstract
A party interacts with a blockchain to notarize his or her IP assets and then uses an unsupervised machine learning pipeline to compare such IP assets to other assets from competitors, as illustrated here with 3D printing datafiles (Lego-like bricks). The table shows similarity scores (in percent) between bricks of a given IP owner (blue) and a brick from a competitor (red). Such an approach can be used by IP professionals to assess the validity of potential IP rights and a possible infringement of such rights.
Introduction
Artificial Intelligence (AI) and blockchain are technologies that are already impacting the world of intellectual property (IP). For example, IP databases now routinely involve machine learning to search trademarks and patent documents [1,2]. More generally, machine learning applications to various IP fields have been reported, including patent monitoring, copyright assessment, and trademark comparisons [3]. Moreover, a wide offer of blockchain solutions are being proposed for the notarization of IP assets [4,5] and new platforms emerge, which leverage both AI and blockchain with the aim to create new IP ecosystems [6,7].
The rise of such technologies in IP is essentially driven by the information technology (IT) industry, in particular tech entrepreneurs [3,4,7]. The available solutions mostly target IP consumers (companies having an active IP policy), allowing them to bypass traditional IP firms (IP law firms, patent and trademark firms) in some cases.
In addition, the law professions seem to be rather slow to adopt “Legal Tech” solutions [8,9]. Understandably enough, IP firms are hesitant about AI and blockchain technologies, with good reasons: a wrong choice of technology may have irreversible consequences, notably for their clients [10]. Plus, seasoned IP practitioners typically have no first-hand experience in these fields, very little time to invest in exploring the proposed solutions, and further regard them as potential threats to their traditional business models. For completeness, the hypes about AI and blockchain gave rise to legitimate questions, but also to some deception, fantasies, and myths [11,12]. In other words, such technologies remain obscure to many IP professionals and are still outside the reach of most IP firms in practice, notwithstanding the current democratization of AI and blockchain.
Does this mean that IP practitioners (lawyers, trademark attorneys, and patent attorneys) are not interested in such matters? No, there is a genuine interest for such tools [13] and a growing proportion of IP professionals specialize in such technologies. A quantity of papers, news, and posts directed to AI or blockchains are being published in the attention of IP practitioners. However, they mostly address the technology in abstract terms. So far, little basic applications have been demonstrated, which pedagogically address the underlying technology and are of direct interest to IP practitioners. What is typically missing in the current literature are use cases exemplifying how such technologies can effectively be exploited in the lifecycle management of IP rights, with basic examples.
This is precisely the main purpose of this paper, which illustrates how blockchain and machine learning can be leveraged in the IP lifecycle management of 3D printing datafiles. Such datafiles are digital representations of 3D objects that may potentially be subject to various IP rights, including copyright, trademarks, design patents, utility patents, and prior use rights. As such, they offer use cases of potential interest to many IP practitioners. Another purpose of this paper is to recount experience and views of IP professionals who have developed and tested AI and blockchain solutions, starting with the pipelines used to perform computations for this paper.
In detail, the next section discusses how AI and blockchain (or other cryptographic solutions) can be leveraged by IP professionals. The third section describes an IP lifecycle management process involving both AI and blockchain, which technologies complement each other well for IP matters. Section four gives an example of application to 3D printing datafiles. The last section closes the paper with some concluding remarks as to pros and cons of such technologies, and how IP law firms may start operationalizing them.
Section snippets
AI vs. machine learning
AI is today a technical reality that is gaining ground in just about every business sector. This logically generates IP rights, such as trademarks and copyright (stemming from computer programs), as well as neighbouring rights, such as open source licenses. Moreover, the innovation pace has caused patent applications to spectacularly increase since 2013 [14].
Generally, AI refers to computerized machines performing tasks that typically used to be associated with cognitive functions of the human
Example of IP lifecycle management process using both machine learning and blockchain
The following describes a process for managing IP rights associated to 3D printing datafiles. More generally, though, the same approach can be applied to any type of datafiles (e.g., images, audio, text, product datasheets, etc.) and thus any kind of IP assets that can be digitally captured as datafiles.
Such datafiles may notably correspond to works eligible to copyright protection. In this case, it is useful to log evidence for the creation of the works, in order to be later able to enforce
Application to 3D printing datafiles
The context is the following: the owner A is assumed to systematically store 3D printing datafiles corresponding to her/his creations: Lego-like bricks in this fictive example. Such bricks may possibly be subject to copyrights, design patents, and utility patents. In principle, a 3D model can potentially be used as a basis for a trademark application too. We note, however, that 3D mark protection was refused in several jurisdictions for Lego bricks [22].
In addition to securing IP rights, the
Conclusion
Clearly, machine learning has potential to help IP practitioners perform IP-related tasks. An appealing application is the measure of similarities between many IP assets or IP rights. Such tasks will become customary in the future, owing to the rapid growth of digital contents (intangible assets already make up the bulk of the S&P 500 market value). It is therefore in the interest of IP professionals to familiarize themselves with such tools and, importantly, to understand the limits of such
CRediT authorship contribution statement
Sébastien Ragot: Conceptualization, Methodology, Formal analysis, Writing - original draft, Writing - review & editing. Antje Rey: Writing - review & editing. Ramin Shafai: Writing - review & editing.
Declaration of competing interest
All of the authors are employed in an Intellectual property law firm. As such, they are independent from their clients. The authors hereby state they have no financial interest, personal interest, or belief that could affect their objectivity.
Acknowledgements
The authors wish to thank Alice Barbet-Massin for her insight into legal implications of blockchains, Gero Dittman and Ian Camenish for fruitful discussions, as well as Martin Toleti and Kurt Sutter for reviewing the manuscript.
Sébastien Ragot is a partner with E. Blum & Co. AG. in Zurich. He holds a PhD. in Materials sciences from Ecole Centrale Paris and is registered as a European Patent Attorney and a Swiss Patent Attorney. He began his career in intellectual property as a patent professional at Hirsch & Partners in Paris, then acted as in-house counsel at IBM Research Europe in Zurich, before joining E. Blum & Co. His legal practice focuses on the drafting and prosecution of patent applications, with core
References (23)
Information retrieval, machine learning, and Natural Language Processing for intellectual property information
World Patent Information
(June 2017)WIPO launches state-of-the-art artificial intelligence-based image search tool for brands
- et al.
WIPO communication
- et al.
Unsupervised learning based patent landscapes using full-text patent data
- et al.
A Survey of Blockchain Technologies for Open Innovation
Blockchain and IP Law: A Match made in Crypto Heaven?
(February 2018)Blockchain Party
Managing Intell. Prop
(2018)IPwe Platform Analytics White Paper
The use and regulation of technology in the legal sector beyond england and wales
Anwaltskanzleien verpassen die digitale Revolution
(July 2018)
Four Areas of Legal Ripe for Disruption by Smart Startups
Cited by (11)
How can Blockchain technology support patent management? A systematic literature review
2023, International Journal of Information ManagementCitation Excerpt :Such issues as high transaction costs, responsibility conflicts, free riding, and the lack of trust between actors are still ubiquitous (Du Chatenier et al., 2010; Lichtenthaler & Ernst, 2006; van de Vrande, Jong, de, Vanhaverbeke, & Rochemont, 2009). Even though traditional notaries provide similar services, they are outperformed by Blockchain notary systems in three respects: These systems enable an automatic recoding and validation for the existence of R&D outcomes, they do not require trust between stakeholders, and they considerably reduce transaction costs for recording and validating the existence of R&D outcomes (e.g. Barulli, 2021; Lin et al., 2020; Maesa & Mori, 2020; Modic et al., 2019; Ragot et al., 2020; Singh & Tripathi, 2019; Vella et al., 2018). Schoenhals et al. (2018) propose a four task approach for time-stamping R&D outcomes (Table 6).
Measuring the originality of intellectual property assets based on estimated inter-asset distances
2022, World Patent InformationCitation Excerpt :Feature extraction algorithms (sometimes called encoders) allow input data to be converted into vectors. Various extraction methods exist, which allow a variety of digital content to be processed, such as images, 3D data, sounds (audio files), and text documents, including product datasheets and patents [15–17]. Next, distances can be simply computed between points mapped by the vectors extracted from input datafiles.
A review on digitalization trends in patent information databases and interrogation tools
2022, World Patent InformationCitation Excerpt :In this vein, Blockchain technology merits mention as an emerging digitalization trend. A Blockchain is a distributed ledger, comprising records of information stored in blocks that are linked by use of cryptography [39–41]. This technology allows storing data in a tamper-resistant way and transferring it securely by means of public-key cryptography.
Uncovering the Complexities of Intellectual Property Management in the era of AI: Insights from a Bibliometric Analysis
2023, Journal of Competitiveness
Sébastien Ragot is a partner with E. Blum & Co. AG. in Zurich. He holds a PhD. in Materials sciences from Ecole Centrale Paris and is registered as a European Patent Attorney and a Swiss Patent Attorney. He began his career in intellectual property as a patent professional at Hirsch & Partners in Paris, then acted as in-house counsel at IBM Research Europe in Zurich, before joining E. Blum & Co. His legal practice focuses on the drafting and prosecution of patent applications, with core technical skills in information technology and physics. He is counseling clients in various intellectual property areas, including open source and software copyright, and has experience with patent infringement and validity, patent litigation, and oppositions and appeals before the European Patent Office. He further develops machine learning solutions for intellectual property..
Antje Rey is a European Patent Attorney and a Swiss Patent Attorney, and a partner with E. Blum & Co. AG. She studied mechanical engineering at ETH Zurich, and completed research activities at IBM Research Europe in Zurich along doctoral studies at ETH Zurich in the field of nanoelectronics and nano-optics. She earned a doctoral degree from ETH Zurich and was a lecturer at the department of Mechanical and Process Engineering of ETH Zurich until 2014. She joined E. Blum & Co in 2014, where she is practicing patent law with core activities in mechanical engineering, micro- and nanotechnology. Her legal practice ranges from the preparation and prosecution of patent applications to counseling of clients in patent-related matters, such as patent infringement and validity, patent litigation, and oppositions and appeals before the European Patent Office.
Ramin Shafai Shafai is a trademark and design specialist, IT application manager, and partner with E. Blum & Co. AG. in Zurich. He studied at the Faculty of Law and the Institute of Computational Linguistics of the University of Zurich. He has a seasoned experience in planning and implementation of internet projects (HTML, CSS, PHP, MySQL), and a comprehensive knowledge in digital image processing and illustration. He joined E. Blum & Co in 2005 and works in the trademarks division of E. Blum & Co since 2012. He is counseling clients in various areas of intellectual property law, including trademarks and design, corporate and trade names, copyright and unfair competition.