AI Unleashed: Navigating Legal Frontiers in Inventorship, Personhood, and Data Dynamics

30th December, 2023 - 11:26 am Categories: AI Patents 0 Comments

Introduction

The growth of technology in today’s era has led to the immense development of Artificial Intelligence (AI), leading to tremendous innovations. AI has become a global phenomenon in terms of its use and is experiencing exponential growth. The shift in the focus has been towards making these machines capable of emulating the thoughts and actions undertaken only by humans. What was once considered a fiction or fantasy decades ago has become a reality.[i] The growth of AI, with its extraordinary capacity to function without human intervention, introduced ethical-legal dilemmas and unleashed numerous challenges. The debate popularized when in the patent application. Dr. Stephen Thaler listed the AI system DABUS (Device Autonomously Bootstrapping Uniform Sensibility) as its inventor.[ii]Interestingly, this listing faced a lot of criticism and, in fact, was rejected by the United States of America (USPTO), the European Patent Office (EPO), and the United Kingdom (UKIPO). However, Australia[iii] and South Africa’s[iv] acceptance of AI as an inventor in their respective domestic patent laws was surprising. In the context of these developments, it becomes highly imperative to address the implications of inventions generated by artificial intelligence (AI) and their legality of patentability, considering their massive impact on the global invention sector and for future discussion.

The present literature on AI has mainly focused on its potential misuse, growth, and liability concerns. However, there needs to be more discussions regarding the necessity for legislation in favour of an open data marketplace to help develop the attribution of full legal personality to AI. Considering the rapid growth of AI inventions, this paper argues, based on the utilitarian concept,[v] that the scope of legal personhood should be broadened and based on those inventions that can solve a particular problem based on necessity rather than relying on theoretical notions of personhood which is the reflection of creator’s personality,[vi] Conventionally, AI has been perceived as a ‘tool,’ creating doubt regarding its recognition as a legal personhood, similar to corporations. It holds the end user or manufacturer accountable for the reason that the AI inventions are the result of intervention by Humans, making them liable for the same. However, this exploration has departed from this conventional understanding of the invention (from the human-centric singular ownership model) to the novel experience of acknowledging every single shareholder. The paper’s approach involves the evaluation of the applicability of recognition based on the proportional contribution of any invention that would result in the shift away from a singular human-centric ownership system.

Defining Artificial Intelligence (AI)

AI is an artificially created intelligence associated with developing technologies that help computers operate intelligently, imitating human-like behaviour. These technologies are different from traditional computer programs in that they assess data autonomously and function without human intervention rather than automatically.[vii] These developments have far crossed the age of calculators to contemporary AI systems; for instance, Google’s self-driving cars, digital assistants, and robot nurses function very delicately and intricately. This exhibition of AI has proved to go beyond mere objects, behaving as independent and distinct entities.[viii] Currently in its developmental stage, AI holds an immense potential to act, function and think human-like intelligence or even surpass human intelligence, reaching the stage of superintelligence.

The intersection of AI and Patent Laws

[Image Sources: Shutterstock]

However, the growth of AI-generated creations has been controversial and poses a challenge to its legality worldwide. In the general context of patent laws, an invention denotes a product or process offering users a novel approach to performing a specific action, potentially providing a new solution to an existing technical problem.[ix] Inventors are granted exclusive rights over their creations, resulting in a monopoly of the owner in terms of the selling power and its usage for a temporary period. While patent laws aim to foster innovation and creativity, the difficulty arises when AI systems generate patentable inventions, going beyond their role as mere tools. This dilemma has ignited debates surrounding AI invention, liability, accountability and ownership. Therefore, it becomes imperative to address these challenges with the help of adaptive legislation and establish comprehensive directions and guidelines for inclusive patent laws.

Contextualizing Indian Patent Laws and AI-Generated Inventions

In India, the intersection of patent laws and Artificial Intelligence (AI) bring a nuanced and complex landscape with implications for inventorship, legal interpretation, and the burgeoning nature of technological advancements.

As per Section 2(1)(j) of the Indian Patent Act, 1970,[x] an “invention” is defined as a new product or process involving an inventive step and capable of industrial application. Section 6 further outlines who is eligible to apply for a patent, encompassing any person claiming to be the first and the true inventor, the assignee of such a person, or the legal representative of a deceased person entitled to make the application.

After reading these provisions, one can find that there has been a complete disregard for AI as an inventor. Often, the debates revolve around the interpretation of the term “true and first inventor”, where it is assumed that the term exclusively deals with a natural person. However, the Act does not expressly limit this to a ‘natural person.’[xi] For instance, mentioning the term’ government’ as a non-natural person under Section 2(s) indicates a broader approach to legislative intention. Judicial perspectives from cases such as Bharat Cooperative Bank (Mumbai) Ltd v. Employees Union in India,[xii] the Australian Patent Act, and South Africa’s Patent Act of 1978 reinforce the argument that “inventor” should not be read narrowly.

However, the provisions of the Indian Patent Act do not expressly mention the threshold up to which human involvement is required, focusing more on the inventive aspect of creation.[xiii] Judicial precedents, for example, the case of V.B Mohammed Ibrahim vs. Alfred Schafranek & Ors,[xiv] emphasized a technical contribution from the inventor. More importantly, legal persons like corporations are not eligible to become ‘inventors’ under the Act. The determination of inventorship becomes complex when considering AI’s training and data consumption, as only some things the AI creates can be definitively traced back to human influence.

Evaluating inventorship scenarios, three possibilities arise;[xv]

1. Inventions of AI and human-made are utilized for the outcome verification.
2. Cases where humans identify a problem and employ AI to find a solution.
3. Instances of AI-generated inventions, where AI independently identifies a problem and proposes a solution without human intervention.

While the first scenario is direct, the second hinges on whether humans provide computational support, resulting in a significant debate on AI-aided invention. The third scenario, where AI operates autonomously, threatens the traditional notions of inventorship.

The contention for the patent applicability of AI becomes more important when its contribution surpasses that of humans[xvi] or when considering the cognitive autonomy of AI. Challenging the status of AI as a ‘true and first’ inventor, more importantly in situations where AI achieves novelty[xvii] and commercial viability, non-obviousness,[xviii] could lead to false claims of inventorship. Identifying and recognizing AI’s potential to stimulate innovation and commercial utility, the non-granting of patents for AI inventions may go against the economic incentives and objectives of the Patents Act.

Legal personhood of AI

A pertinent question worth asking is whether AI entities can be considered legal persons. Conventionally, legal personhood has been closely associated with humanity,[xix] with philosophers arguing that artificial intelligence lacks the ethical and moral considerations that are inherent in humans and, hence, can only be considered as slaves[xx] or tools[xxi] of humans. On the Contrary, Dyschkant argues that we should be focusing on the technical definition of a ‘person’ subject to legal rights and duties while divorcing the concept of humanity from legal personhood, extending beyond natural persons to include non-human beings such as machines.

The concepts of persona ficta[xxii] and legal person have been suggested to include legal personhood for AI. This would mean recognizing AI as a fictitious person with obligated duties, a persona ficta with recognized will (either attributed from a human or inherent), or a legal person without the need for a choice. AI, often considered a ‘Weak AI,’ dependent on humans for data and programming, lacks the autonomous characteristic of ‘Strong AI.’ As a result, the conferment of legal rights to AI is contingent upon ownership by humans.

Concerning duties, it is contended that entities unable to fulfil their obligations cannot be classified as legal personhood.[xxiii] If AI is accepted as an artificial person, it might need to possess assets, own property, or generate revenue, forming the basis for imposing liability. The intrinsic connection between ownership and legal personhood[xxiv] becomes relevant in evaluating these considerations.

Ownership and Liability of AI-generated Inventions

Ownership and liability of AI-generated inventions push further for a voyage on the evolution of proper and legal personhood concepts. If the law takes into consideration something mere property, it would lack legal personality; in contrast, if it owns property, it will possess legal character. This is evident in recognizing corporations and rivers as legal persons. While the right to own property is given to corporations, they are ultimately for the human interest. Likewise, it is contended that AI, despite increasing autonomy, cannot own property without being dependent on humans; the latter’s presence is essential as AI systems remain controlled and programmed by humans.

AI is currently categorized as’ dependent on human input and lacking full autonomy. AI’s incapacity to enter contracts, respond to financial incentives, avoid harm, initiate legal proceedings, or authorize license independently highlights its reliance on human guidance.[xxv] Addressing potential violations of laws, particularly under the Indian Patent Act, necessitates a procedure to determine liability in situations where AI technology infringes patent claims.

Product liability principles can also apply to AI inventions, emphasizing defects, manufacturing defects, and failure to warn against non-obvious threats.[xxvi] The intervention of humans in the AI process, a characteristic of Weak AI, allows for distinguishing between AI’s autonomous decisions and damages from product defects. According to the European Parliament Resolution, liability for patent infringement is upon legal entities and natural persons involved in the development of AI, holding them accountable.[xxvii]

The accountability of AI developers is nothing but a human-centric approach. However, as AI grow, singular systems may create multiple inventions and assess patent risks, involving various stakeholders such as programmers, trainers, owners, data suppliers and end-users. A comprehensive evaluation of AI accountability necessitates careful consideration. Analogous to corporate entities, AI liability arises proportional to the autonomy and learning capability, recognizing the responsibility of trainers.[xxviii]

Contending for a liability concept akin to a corporate structure, where stakeholders have different interests and roles, requires a subtle approach. A contractual framework specifying each stakeholder’s purposes, liabilities, rights, duties, and contributions accommodates the diversities in AI entities. This approach recognizes the various natures of AI involvement and determines liability based on the capacity of each stakeholder. Incorporating AI as a legal person, alongside existing patents, acknowledges its status without mandating legal entity registration for all patented AIs.

Necessity of granting legal personhood to AI

The necessity of recognizing AI as a legal personhood is grounded in legal principles as well as the practical considerations. Salmond proposes that, any being capable of rights or duties, irrespective of being a human or not, can be considered as a person.[xxix] While the UKIPO acknowledges the lack of legislative clarification, renders it difficult to recognize AI entities as a legal person. In Indian jurisprudence, entities like idols,[xxx] animals[xxxi] and rivers[xxxii] have been recognized as personhood, has historically arisen through judicial innovation rather than legislative action.

Granting legal personhood to AI improves the legal efficiency. Analogous to corporate liabilities, recognizing AI entities as legal subjects would allow anyone to sue wrongdoings of AI as well protecting the rights and duties of all, and streamlining processes that would otherwise involve identifying individual human actors behind the AI inventions, leading more consumption of resources and difficulty in finding justice on time.[xxxiii] In situations involving transactions between multiple AI systems, clarification with respect to assigning liability poses a challenge, without recognizing AI as subjects. This recognition addresses the intricacies of responsibility and facilitates smoother operations.

In the realm of patent law, ambiguity surrounding the patentability of AI-based inventions in India underscores the need for legal personhood. The legal framework needs more clarity on accommodating AI as inventors, posing challenges for rights holders. Recognizing AI as legal persons subject to rights and duties empowers owners to apply for patents and act on behalf of the AI. This ensures the owner’s enjoyment and aligns with the objectives of the Patents Act.[xxxiv]

Legal personhood is justified for weak AI through human guardianship, establishing constructive ownership. This parallels the legal status of corporations and idols, wherein control and possession are exercised through human agency. As discussions about the personhood of strong AI emerge, the complexities of their autonomy warrant careful examination. The unpredictability of strong AI raises questions about potential property rights and the extent of their independence beyond human control. While legal recognition for strong AI may be subject to future developments, the present case for weak AI emphasizes the need for legal personhood based on existing legal analogies and practical considerations.

AI as a Legal Entity and Data Monopolisation

In the Indian context, the regulation of data is relatively underdeveloped.[xxxv] AI invention relies heavily on data consumption for machine learning, making the data provider a crucial stakeholder in the contractual model, similar to corporations discussed earlier. The absence of robust legislation on data consumption creates an environment where big-tech companies possessing extensive data repositories can assert dominance in AI. This dominance leads to a scenario where such companies have stakes in multiple AI entities, resulting in a monopolistic trend. The dependence of AI on data sets establishes a positive feedback loop, contributing to the concentration of power among major players.

Critics argue that the benefits of innovation outweigh the potential negative consequences of monopolies, such as market imbalances. However, this perspective overlooks the harmful effects of intellectual property monopolies and their impact on market innovation. Big-tech companies further perpetuate their dominance by absorbing smaller competitors through corporate acquisitions, creating a self-reinforcing cycle of monopolization. Recognizing the significance of open data marketplaces, the National Strategy for Artificial Intelligence in India suggests making specific government data available for the ‘public good.’[xxxvi] The 2018-2019 Economic Survey of India considers data a natural resource, emphasizing that anonymized personal data transforms into a ‘public good.’ Access to these data sets has the potential to benefit the private sector in the development of AI and AI-generated inventions. Currently, the private sector in India needs help to realize the societal benefits of available data.[xxxvii]

Conclusion

The growth of Artificial Intelligence (AI) has led to complex legal challenges, starting from inventor ship debates to the recognition of AI as a legal persona. This exploration has traversed from a human-centric ownership model to a stakeholder-based approach regarding liability. Highlighting the necessity for recognizing AI as legal personhood, the paper navigates the complexities of patent laws, emphasizing the potential hindrances in India’s current legal framework. This investigation underscores the importance of open data marketplaces addressing data monopolization risks. Recognizing AI as a legal entity alongside existing patents is proposed, acknowledging its diverse nature. Overall, adapting legal frameworks is essential to balance innovation and ethical-legal considerations in the burgeoning field of AI.

Author: Shivam Sharma, in case of any queries please contact/write back to us via email chhavi@khuranaandkhurana.com or at Khurana & Khurana, Advocates and IP Attorney.

[i]John Haugeland, Artificial Intelligence: The Very Idea (MIT Press, Cambridge, MA 1985).

[ii]IPWatchdog, DABUS Gets Its First Patent in South Africa Under Formalities Examination, IPWatchdog, (July 29, 2021), https://www.ipwatchdog.com/2021/07/29/dabus-gets-first-patent-south-africa-formalities-examination/id=136116/.

[iii]Donrich Willem Thaldarm & Meshandren Naidoo, AI: Inventorship: The Right Decision?, OSF Preprints, (Aug 4, 2021), https://osf.io/7uctg/.

[iv]Ibid

[v]Justin Hughes, The Philosophy of Intellectual Property, 77 GEO. L.J. 287  (1988)

[vi]3 G.W.F. HEGEL, PHILOSOPHY OF RIGHT (S.W. Dyde trans., Prometheus Books 1996) (1821)

[vii]Paulius Cerka et al., Is it possible to grant legal personality to artificial intelligence software systems?, Computer Law & Security Review, 33(5), 685–699.

[viii]Marshal S.Willick Constitutional Law and Artificial Intelligence: The Potential Legal Recognition of Computers as “Persons”, IJCAI (1985).

[ix]WIPO, Patents, https://www.wipo.int/patents/en.

[x]The Patents Act, 1970, No, 39, Acts of Parliament, 1970 (India).

[xi]Anmol Maheshwari, Dawn of Artificial Intelligence Changing the Face of Patent Regime, 5 Amity International Journal of Juridical Sciences, 131, 126-135 (2019).

[xii]4 SCC 685 (2007).

[xiii]Pankaj Soni & Kartikay Vikrant Singh, Are we ready for AI disruption? An Indian patent law perspective, Remfry & Sagar, https://www.remfry.com/wp-content/uploads/2018/10/Print-PDF.pdf.

[xiv]AIR Mys 173 (1960) (India).

[xv]European Patent Office [EPO], Update of legal aspects of artificial intelligence and patents, EPO doc. CA/PL 5/20 e (2020), https://www.epo.org/modules/epoweb/acdocument/epoweb2/468/en/CA-PL_5-20_en.pdf.

[xvi]Daria Kim, ‘AI- Generated Inventions’: Time to Get the Record Straight?, GRUR International 69:5, 443-456 (2020).

[xvii]The Patents Act, 1970 § 2(1)(l)

[xviii]The Patents Act, 1970 § 2(1)(ja)

[xix]Alexis Dyschkant, Legal Personhood: How We Are Getting It Wrong, U Illinois L Rev, 2075- 2109 (2015).

[xx]Joanna J. Bryson, Robots Should Be Slaves’ in Yorick Wilks, in Close Engagements with Artificial Companions 63-74 (John Benjamins Publishing Company 2010).

[xxi]Ian R. Kerr, ‘Ensuring the Success of Contract Formation in Agent Mediated Electronic Commerce 1 Electronic Commerce Research journal, 183-202 (2001).

[xxii]G Deiser, The Juristic Person 48 U Pennsylvania L Rev 131, 133 (1908).

[xxiii]People ex rel Nonhuman Rights Project, Inc v Lavery, 2014 WL 6802767 (N.Y. App. Div. Dec. 4, 2014).

[xxiv]Rafael Dean Brown, Property ownership and the legal personhood of artificial intelligence, Information & Communications Technology, 30:2, 208-234 (1964).

[xxv]LexOrbis, Issues in Recognition of Artificial Intelligent Entity as Inventor, (Jun 1, 2020)https://www.lexorbis.com/issues-in-recognition-of-artificial-intelligent-entity-as-inventor/.

[xxvi]Anindya Ghosh & Nabarun Chandra Ray,  India: Product Liability in India: An Evolution, Mondaq, (Sept. 10, 2021, 12:00 AM), https://www.mondaq.com/india/dodd-frank-consumer-protection-act/974270/product-liability-law-in-india-an-evolution.

[xxvii]Dr. Shlomit Yanisky Ravid & Xiaoqiong Liu, When Artificial Intelligence Systems Produce Inventions: An Alternative Model For Patent Law at the 3A Era, Cardozo Law Review, 39, 2215-2262, (2018).

[xxviii]Draft Report with recommendations to the Commission on Civil Law Rules on Robotics, European Parliament (2014-2019), http:// www. europarl.europa.eu/sides/getDoc.do?pubRef=-//EP// NON SG M L%2BCOMPARL%2BPE-582.443%2B01 %2BDOC%2 BPDF% 2BV0 //EN.

[xxix]John Salmond, Jurisprudence 298 and 299 (PJ Fitzgerald ed, 12th ed. India reprint, Thomson Sweet & Maxwell, Universal Law Publishing Co 1966).

[xxx]Indian Young Lawyers Association v State of Kerala 11 SCC 1 (2019).

[xxxi]Karnail Singh v State of Haryana, SCC Online P&H 704 (2019) (India).

[xxxii]Mohd. Salim v State of Uttarakhand, PIL No. 126 of 2014 (India).

[xxxiii]Bendert Zevenbergen et al., Appropriateness and Feasibility of Legal Personhood for AI Systems, ICRES 59-64 (2018).

[xxxiv]Rohan Deshpande & Karan Kamath, Patentability of inventions created by AI- the DABUS claims from an Indian Perspective, Journal of Intellectual Property Law & Practice, 15: 11, 879-889 (2020).

[xxxv]Tim Dornis, Artificial Intelligence and Innovation: The End of Patent Law as We Know It, Yale Journal of Law & Technology, 23(Fall), 98–159, (2021).

[xxxvi]Niti Aayog, Responsible for AI #AIFORALL: Approach Document for India Part 1- Principles for Responsible AI , (February 2021).

[xxxvii]Economic Survey 2018-2019, Data “Of the People, By the People, For the People”.