Abstract
The inception of artificial intelligence has tapped into the sector of ‘innovation and creativity’ and led to the formation of ‘Inventive AI’ to create and develop inventions and works. This has led to doctrinal challenges in the IPR regime wherein ‘only natural persons or legal bodies’ are considered as inventors under patent laws, since the formulation of IPR laws and doctrines which at that time must not have envisioned the scenario of ‘Inventive AI’. In current times, debates, deliberation and discourse must be conducted to reason as to whether to include ‘autonomous AI systems’ as inventors. The blog aims to discuss the aforementioned in view of the recent rejection of a patent application wherein the applicant filed for an ‘AI system’ to be given the designation of inventor for creation of the invention enlisted in the patent application filed before the European Patent Office.
Introduction
In an age where there are seeming observations of the rise of ‘the fourth industrial revolution,’ there arises a need to take into consideration the ‘emerging aspects’ of patent law. These have to constantly keep up with the dynamic advancements in various sectors of innovation more precisely, technological advancements. Underpinning this, the WIPO Technology Trends Study probes that there is a recent global rise in ‘AI Inventive Activity’ wherein IBM and Microsoft, both tech-giants are at the forefront[i].
The notion of artificial intelligence as transformative or disruptive in nature has opened up a Pandora’s Box of complexities. However, with fast-paced technology, the transformative aspect of AI has outmatched the disruptive nature of the technology by a fine line[ii].
Transformative Artificial Intelligence has opened up discourse and debate regarding inventions conceived by ‘AI systems’ (can be termed as Inventive AI) and the supplementary inventor-ship issue that goes along with it[iii]. Inventive AI can most closely be associated with inventions wherein AI systems have majorly contributed to the development or formation of the invention. To put it in technical terms, it is the AI systems that implement ‘the inventive step’[iv].
The consideration of an ‘AI system’ as ‘an inventor’ is indistinctly worded for unpredictable technological advances in the future. The exact scope of which is and will remain debated and criticised until the code is cracked.
Artificial Intelligence Systems as an Investor
In contemporary society, the application of AI has left no strata of any domain untouched. Thereby, inadequate preparation of contemporary patent laws vis-a-vis ‘Inventive AI’ will result in ineffective and outdated patent laws. This shall inevitably have a ripple effect on highly advanced technological innovations generated and the consequent negative effects on:
Innovation
Incentive and rewards
Sufficient disclosure of claimed invention
Commercialization and development of inventions.
The resulting outcome of which is ‘stifled growth and development’[v]. Revisiting patent laws in light of advancement of technologies to overcome challenges such as the ‘test of skilled person in the art’ juxtaposed to Inventive AI wherein considerations for the bar of patentability will have to be raised, are one of the many pitfalls that will arise.[vi]
One of such challenges is the issue of patent application filings wherein ‘an AI machine’ is to be designated as the inventor. This is another legitimate cause to re-evaluate patent laws bearing in mind the inception of ‘Inventive AI’.
DABUS Patent Application
In the month of November 2019, a patent application was filed before the European Patent Office[vii]. Herein, the applicant in his application asserted the following:
The machine called ‘DABUS’ is the brainchild of the Dr. Stephen Thaler of ‘Imagination Engines’[x] supported by a team of international patent attorneys constituted under ‘The Artificial Inventor Projector’. DABUS machine conceived two inventions; one, a ‘food container’ and second a ‘device and method for attracting enhanced attention’ autonomously without any form of human intervention.[xi]
Decision by the European Patent Office
The decision of the European Patent Office was released a month later rejecting the aforesaid patent applications on the grounds of non-compliance with Article 81 ‘Designation of Inventor’[xii] and Rule 19 ‘Designation of the Inventor’[xiii] of the European Patent Convention. To be more specific, the office asserted that the existing legal framework of intellectual property is applicable only to natural persons, legal persons or bodies’ equivalent to legal persons. Evidently ‘Artificial Intelligence’ does not fall within any of the aforementioned categories.
The judgement in brief, dwelled on the concept of ‘legal personality’ and how it does not apply to ‘AI Systems’ as it has not been attributed the same by virtue of any legislation or jurisprudence. Furthermore, the judgement also outlined that ‘AI Systems’ by way of having no ‘legal personality’, would have no rights therefore they cannot have legal title over their output or invention or transfer it or even be employed for the sole reason of absence of rights which flows from legal personality.[xiv]
Submissions made by the aforesaid patent applicant and several others advocating for ‘AI systems’ to be given recognition of ‘e-personhood’ put forth that the current AI ecosystem comprises of ‘sophisticated autonomous AI systems’. In the case of ‘DABUS’, the applicant contended that the machine itself conceived the idea, and materially contributed to its implementation and development with no form of human intervention. Not acknowledging DABUS as an inventor would mislead the public in as to who is the actual devisor of the invention.[xv]
The case of ‘DABUS’ has opened up discourse and debate in the unchartered territory of ‘grant of personhood to AI systems’[xvi], with certain jurisdictions such as the EU already considering ‘Civil Law Rules On Robotics’.[xvii]
Overview of the Current Jurisprudence and Legislative Framework
In regards to patent law, the current jurisprudence and legal framework speaks for itself when it asserts that ‘it does not recognize AI systems or machines as inventors and only natural persons or legal persons’. Thereby, underscoring the anthropocentric view it encapsulates. [xviii]
The genesis of patent law was formulated to spur innovation and progress by humans in exchange for incentive and reward. However, the current view stands justified only to a certain extent particularly due to the rise of inventive artificial intelligence.
Enormous advancements and progress in AI technology has caused the need for re-consideration and revisions concerning legal philosophical questions on the incentive theory of the patent system as inventions are being conceived by machines which lack the desire for rewards or incentive.[xix]
Categories of Artificial Intelligence Inventions
As maintained by several reports published by international organizations and forums such as WIPO[xx], IP5 organizations[xxi] with respect to AI, inventions can be categorized as those wherein AI is used as an assisting tool in human created inventions, incorporating AI methods in creating inventions and lastly inventions generated by AI autonomously with no form of human intervention.
The third categorization has opened a whole discourse on the feasibility of considering ‘sophisticated autonomous AI systems’ as an inventor. This premise clearly indicates that the current jurisprudence and legislation in patent law requires revisions.
Advocates, supporting the novel idea of ‘e-personhood’, maintain that in the current law with regards to inventor-ship there is discrimination. The ‘designation of an inventor’ will be conferred upon natural person, such as the creator or owner of the machine due to the present legal framework’s non-recognition of anything other than natural persons or legal bodies as ‘inventors’, as opposed to the actual devisor of the machine which is an ‘autonomous AI system’.
Classic example of the aforementioned is the rejection of the DABUS patent application previously mentioned. It must therefore entail that proposals of laws and changes must be made such that that fairness with regard to creation of invention and innovation go hand in hand.
Particularly in regards to patent law, since Inventive AI creations are on the rise and some of which such as DABUS are autonomously generating inventions sans human intervention. The idea of ‘conception’ in the mind of the inventor and subsequent implementation of the conception is one of the main grounds for patent filing.[xxii] But if the idea itself is conceived by a machine and not by a ‘natural person,’ the basic norm will be challenged which is bound to give rise to doctrinal challenges of traditional intellectual property laws.
Case for Grant of 'E-Personhood' to Machines and Robots
At present the field remains divided as between IPR experts and jurists on the issue of granting legal personhood to ‘sophisticated autonomous AI systems’, as fundamental jurisprudential concepts will need to be revised in light of the given situation.
Individuals in support of ‘e-personhood’ base their stand on several grounds. Some of them put forth that systems such as autonomous AI are already beginning to steadily resemble primary skills set and behaviour of human beings, such as autonomy in making decisions, self-learning and adaptability, interactiveness, efficiency ,etc. Advocates in the field of intellectual property maintain, that in the present schemes of things AI systems are autonomously generating inventions and works while simultaneously fulfilling the tests of patent and copyright law. The former test is mostly associated with novelty[xxiii] and the latter test with creativity[xxiv].
Fuelled by AI, examples including Philyra by IBM that creates and develops new fragrance formulas[xxv] which otherwise requires scrupulous human skills; Dr Watson by IBM that is able to conduct patient diagnosis on the basis of the data-set it is fed[xxvi]; GAN system aided by AI created portraits of fictitious individuals which was later sold at Christie’s for an astounding amount[xxvii].This list of creative and inventive AI is not exhaustive and can continue for several more pages. Applications of AI by degrees will inevitably result in the development of autonomous and self-adaptable sophisticated systems.
Intellectual Property Scholars are propounding legal models such as ‘partial legal status to be attributed to autonomous AI systems’[xxviii] ‘establishing the concept of principal-agent vis-à-vis human principal-AI system’[xxix] so as to explore pragmatic solutions for the upcoming age of innovation.
While dissenters, who believe that ‘e-personhood’ should not be granted contend that there exists no such legal model from where AI systems can derive their personhood. Further contending that the proposal of ‘e-personhood’ must be viewed through the lens of societal ,ethical as well as psychological impacts and not just the economic and legal aspects. Another issue which individuals against grant of ‘e-personhood’ bring up is that of the liability of actions committed by autonomous AI system, for example autonomous driving vehicles. Would the granting of such ‘e-personhood’ absolve any or all form of human liability be it tortious or contractual are some of the concerns that are raised by the opposing parties.[xxx]
It is absolutely justified to construe that the issue of whether ‘AI system being considered as an inventor’ will only be given due deliberation when the current legislative and juristic regimes reform the majoritarian view that an inventor should be a ‘natural person’.
The filing of the DABUS patent application has thereby called into question and kicked off the discourse on the need for traditional intellectual property law doctrines’ human-centric view to be changed. When the doctrinal and legislative views of anthropocentricity are given a consideration for revision, there are bound to be other cascading effects in the IPR regime, for instance tests for copyright and patentability can be considered as cases in point.
The present situation can be considered as between the devil and the deep blue sea. However, to be in tune with progress and advancements, laws must be dynamic and be formulated as such to maximise innovation and incentive especially in the domain of intellectual property where the former and the latter prevail.
[i] WIPO’s ‘First Technology Trends’ Study Probes Artificial Intelligence :IBM and Microsoft are Leaders Amid Recent Global Upsurge in AI Inventive Activity, Available at https://www.wipo.int/pressroom/en/articles/2019/article_0001.html
[ii] Tech Trends 2020: Moving from Disruption to Transformation, Available at https://www.forbes.com/sites/forbestechcouncil/2019/12/09/tech-trends-2020-moving-from-disruption-to-transformation/#6c397a633663
[iii] White Paper on ‘Artificial Intelligence Collides with Patent Law’ , Published by World Economic Forum (April 2018) , Available at http://www3.weforum.org/docs/WEF_48540_WP_End_of_Innovation_Protecting_Patent_Law.pdf
[iv]Inventive AI: Can Machines Innovate?, Available at https://brite.nridigital.com/brite_autumn19/artificial_intelligence_invention_intellectual_property
[v] Supra [iii]
[vi] Ryan Abbott, ‘Everything is Obvious’ 66 UCLA L. Rev. 2 (2019).
[vii] Patent Applications filed before EPO numbered ‘(EP 18275163) and (EP 18275174) , Available at http://artificialinventor.com/patent-applications/ and http://ipkitten.blogspot.com/2019/12/epo-refuses-ai-inventor-applications-in.html
[viii] Ibid.
[ix] Supra [vii]
[x] What is DABUS? , Available at http://imagination-engines.com/iei_dabus.php
[xi] Artificial Inventors, Available at http://artificialinventor.com/dabus/
[xii] Article 81 ‘Designation of the Inventor’ of the European Patent Convention , Available at https://www.epo.org/law-practice/legal-texts/html/epc/2016/e/ar81.html
[xiii] Rule 19 ‘Designation of the Inventor’ of the European Patent Convention , Available at https://www.epo.org/law-practice/legal-texts/html/epc/2016/e/r19.html
[xiv] EPO publishes grounds for its decision to refuse two patent applications naming a machine as an inventor (28 January 2020) , Available at https://www.epo.org/news-issues/news/2020/20200128.html and https://register.epo.org/application?documentId=E4B63SD62191498&number=EP18275163&lng=en&npl=false (Grounds for decision)
[xv] ‘EPO refuses “AI inventor” applications in short order – AI Inventor team intend to appeal’, Minutes of the Oral Proceedings before the EPO (22 December 2019) , Available at http://ipkitten.blogspot.com/2019/12/epo-refuses-ai-inventor-applications-in.html
[xvi] The timeline of ‘e-personhood’: a hasty assumption or a realistic challenge? (25 April 2019 ) , Available at https://www.maastrichtuniversity.nl/blog/2019/04/timeline-e-personhood-hasty-assumption-or-realistic-challenge
[xvii] Draft Report with recommendations to the Commission on Civil Law Rules on Robotics (31 May 2016) , Available at http://www.europarl.europa.eu/doceo/document/JURI-PR-582443_EN.pdf?redirect
[xviii] Supra [xiv]
[xix] Theories of Intellectual Property by William Fisher ,
[xx] Standing Committee on the Law of Patents by WIPO ,(28 May 2019) , Available at https://www.wipo.int/edocs/mdocs/scp/en/scp_30/scp_30_5.pdf
[xxi] Report from the IP5 Expert Roundtable on Artificial Intelligence ( 31 October 2018 ), Available at https://www.fiveipoffices.org/material/ai_roundtable_2018_report/ai_roundtable_2018_report
[xxii] Patent Ownership Vs. Inventor-ship : Who Really Controls the Rights To A Patent ? by Michael Henry (14 June 2018 ) , Available at https://www.henrypatentfirm.com/blog/patent-ownership-vs-inventorship
[xxiii] Novelty as an Invention Patentability Criterion by Iryna Ortynska , Available at https://ipstyle.net/wp-content/uploads/2018/10/Patents.pdf
[xxiv] Howard B. Abrams, ‘Originality and Creativity in Copyright Law’, 55 Law & Contemp. Probs. 3 (1992).
[xxv] ‘Using AI to create new fragrances’ by IBM Research Editorial Staff (23 October 2018) , Available at https://www.ibm.com/blogs/research/2018/10/ai-fragrances/
[xxvi] Industry Report –Patenting in the age of supercomputers by Sumit Prasad (12 July 2017) , Available at https://www.iam-media.com/patenting-age-supercomputers ( In regards to the functioning of Dr. Watson by IBM)
[xxvii] AI Art at Christie’s Sells for $432,500 by Gabe Cohn (25 October 2018) , Available at https://www.nytimes.com/2018/10/25/arts/design/ai-art-sold-christies.html
[xxviii] Thomas. Wischmeyer and Timo Rademacher (eds.), ‘Regulating Artificial Intelligence’ (2020).
[xxix] Digital-Personhood? The Status of Autonomous Software Agents in Private Law by Gunther Teubner (2018),
Available at https://www.anci.ch/
[xxx] Open Letter to the European Commission ‘Artificial Intelligence and Robotics’ , Available at https://g8fip1kplyr33r3krz5b97d1-wpengine.netdna-ssl.com/wp content/uploads/2018/04/RoboticsOpenLetter.pdf
By Karen Baretto, 3rd year student at Kirit P. Mehta School of Law, NMIMS University, Mumbai. This blog is a part of the RSRR Rolling Blog Series.