Methodological Approach to Separating Metaphysical Personhood from a Legal One - AI Perspective
“And if my choice is to sit graciously in my best robes and accept the inevitable or to bail a sea with a bucket, give me the bucket.” (Chalice by Robin McKinley)
Introduction
In this analysis the author argues that psychological and metaphysical characteristics do not afford requisite basis for according legal personhood to Artificial Intelligence (AI). She disagrees with the argument that philosophical personality is a prerequisite for legal one. If this was the case, legal systems would dismiss not only legal personhood of non-humans (e.g.: corporations) or inanimate objects (e.g.: vessels), but also psychopaths, whose metaphysical personality, as opposed to humanity, is debatable. The author concedes that in legal reality it is irrelevant whether AI is a sentient being having full range of emotions or mental capabilities or not.
As suggested by Dewey, as long as the entity in question ‘creates consequences social in character' it may be established as a legal person. The conception one may superimpose on Dewey’s is that there are legal systems that have granted legal personhood to inanimate entities such as a corporation is due to an understanding that the corporation is human association. Contrary, AI, to which this study aims to attribute legal rights and obligations, has no human components. This might be true. Yet, the author, takes a different stand.
She suggests that if only an inanimate object, such as AI creates relations with humans that brings about economical & legal consequences its likeliness of being granted separate legal personhood will be high. With this in mind, this study, aims to bring some order and clarity to the understanding of what legal personhood of AI isolated from philosophical and anthropocentric perspective stands for. Since, there is a close relation between legal and philosophical perspectives this study aims to minimize the risk of disorientation by employing a system of methods adopted to analyse AI as having the potential for legal capacity.
There are at least three methodologies that the author considers to argue AI’s personhood capability. These are: Entity-Centric, Effect/Consequences and Conditions – Based Methodologies. Part II of this analysis will explore advantages & disadvantages of the first methodology. Part III will examine the concept, advantages & disadvantages of the second one. Parts IV will address the advantages & disadvantages of the third one and Part V will conclude the findings.
Entity-Centric Methodology
Advantages & Disadvantages
Entity-Centric Methodology takes into consideration the characteristics of the AI and puts a question: What attributes should AI have in order to qualify as a legal person? This methodology relies on a drawing of analogies with the main features of a currently recognized legal person, namely a corporation. It is the most intuitive approach, where all the rights and responsibilities have been accorded to an entity that fits well within the image of a legal but not natural person, which is a corporation. It creates a figurative boundary between the characteristics of philosophical and legal personality by indicating that ‘all that is necessary for the existence of the person is the lawmaker.
Given the advantages of the entity-centric approach it is important to note some of it inherent disadvantages. First, one could argue that his methodology creates a risk to human distinctiveness. Why this is a problem? For example, the US Supreme Court has extended constitutional rights to corporations. One area in which the US Supreme Court finds it easy to attribute natural person-like qualities to a corporation is constitutional protection of freedom of speech. The attribution of this right to a corporation has been questioned by many scholars and legal practitioners (Atkinson, Glasner, Lock) and their voices employed as a weapon of defense against affording any rights whatsoever to corporations (Clements, Nace, Johnston).
The question then arises about whether AI and its legal personhood would create a similar voice of discontent? The legal personhood is a tool employed by the courts to create or grant rights they assume appropriate and is used both for convenience and to serve the ends of justice. (Wormser) Hence, the type of rights granted to AI would depend on the judiciaryis creating laws, a function that would require them to balance the incentives.
One should keep in mind, however, that entity-centric methodology and corporate legal personality arising from it has been developed without courts providing a clear definition of a ‘person.’ The courts have simply bypassed this preliminary question. To overcome this inconvenience, corporate theorists have devised various interpretations or metaphorical descriptions providing a legal definition for the corporate entity and its personhood. Nonetheless, they never reached an agreement on the one and proper characterization of the corporate entity. This leaves as with various interpretation of the meaning behind the inanimate legal person but with no definition that could be appropriate re-employed with regard to AI.
On second thought this irony is particularly noteworthy as it leaves us quite free to interpret and choose our own conclusions.
Effect/Consequences Based Methodology
Advantages & Disadvantages
Broad suggests, ‘what really matters to science is not the inner nature of objects but their mutual relations will answer all scientific purposes as well as any other set with the same sort of relations,’ Another way of testing AI’s capability of holding legal personhood rights and obligations could be by asking: What does it do? What effects of imposing legal rights and responsibilities on AI have to be evident for AI to qualify as a legal person? Following Broad, the definition of a legal personhood would be an analysis of that which surrounds the entity’s relations and facts, not a search for an inherent essence.
The reason that a molecule conducting electricity is not a legal person, whereas a corporate is, is not that the molecule does not exhibit adequate characteristics analogous to a corporation, but rather that the molecules will not exhibit necessary consequences of acting in such a way as to have an effect on the law. The law conception of the molecule is that it will not change its behaviour.
Dan-Cohen, however, while contributing his voice to this discussion, recognises that there is need of a ‘pre-legal conception’ of the entity before one can evaluate or estimate its nature or legal abilities. One cannot treat ignorantly the nature of the entity , such as a molecule, when its behavioural reactions are relevant. One needs to acknowledge whether the entity is capable of responding to legal stimuli and, if not, how it can affect legal relations and/or engage with them. Consequently, a molecule cannot become a legal person because it will not be able to react to legal stimuli. Will AI be able to react to legal stimuli?
But then again, as it was in respect of the Entity-Centric Methodology, does not it mean that according legal personhood to AI would diminish the human position? As one may see effect/consequence-based methodology does not provide sufficient answer to this question. There is a need for another level of conditions in order to determine the ability for granting undoubted legal personhood to AI. For example, establishing AI in a similar manner as a corporation and providing them with ethical coordinates, such as e.g. – three laws of robotics’ First Law: A robot may not injure a human being, or, through inaction, allow a human being to come to harm. Second Law: A robot must obey the orders given it by human beings except where such orders would conflict with the First Law. Third Law: A robot must protect its own existence as long as such protection does not conflict with the First or Second Law. These laws could serve as conditions under which legal systems would treat AIs as persons.
Conditions Based Methodology
Advantages & Disadvantages
When considering the last theory, namely the conditions-based methodology one needs to ask another set of basic questions, such as: Under what conditions will the law treat AIs as persons? Why is that we need to treat some entities as legal persons and yet not others. Why provide legal personhood to AIs ? After providing answers to the above set of questions adapting the condition-based methodology might provide some advantages. This is, for example, demonstrating that the various entities, such as humans and non-humans, individuals and associations capability of fulfilling the role of becoming a structure of legal logic not metaphysical characteristics.
It also marries well with the condition of granting legal personhood to inanimate objects only when based on the interests of humans and where categorization as a juridical person is necessary for practical reasons, since the law requires an object upon which to act. It underlines one of the main conditions for granting legal personhood to an entity capable of taking responsibility for actions performed autonomously. Oh irony, this utilitarian approach of the condition-based methodology, for some, is its downfall.
Sunstein and Posner are of the opinion that granting personhood is to extend protection to a previously unprotected class of entities. Likewise, Mayer, who suggests that ‘equality of constitutional rights plus an equality of legislated and de facto powers leads inexorably to the supremacy of artificial over real persons.’ As such, transferring rights and duties to inanimate objects risks cheapening human personhood. In the manner previously mentioned, the author disagrees.
If one assumes the distinction between the physical human organism alone and the legal person that is not associated with that organism, only with a legal status granted to it then one must agree that it is permissible to grant legal status to an inanimate object as long as the rights given to it are distinct from human rights. The position here is that human rights attache by the simple fact of being a human being and only the grant of human rights to AI could diminish the human position.
Conclusion
Following the argument of Judge Breitel who pointed out that: [I]t is not true (…) that the legal order necessarily corresponds to the natural order (…); it is a policy determination whether legal personality should attach and not a question of biological or 'natural correspondence.' The conception that this analysis superimposes on this statement and three methodologies listed above is that there exists a way of arguing a vivid picture of AI's potential for legal personhood. The author recognizes that this argument invites inquiry about the type of rights and responsibilities AI should be granted. She concludes that this solution would need to have one aim: to strengthen the legal status of an entity that has real value to the legal world, namely a human being.
References:
(Oscola style of referencing)
Primary sources:
Santa Clara County v Southern Pacific Railroad Company 118 US 394, 396 (1886).
Trustees of Dartmouth College v Woodward 17 US 518 (1819) 636.
United States v The Little Charles [1818] 26F Cas 979, 981-82.
Secondary sources:
J Berg, ‘Of Elephants and Embryos: A Proposed Framework for Legal Personhood’ (2007-2008) 59 Hastings LJ 369, 382-384.
M Del Mar and W Twining, Legal Fictions in Theory and Practice (Springer 2015) 96.
CR Sunstein, 'On the expressive function of law' (1996) 144 University of Pennsylvania Law Review 2021–53.
E Posner, Law and social norms (Harvard University Press 2000).
R Posner, 'The jurisprudence of skepticism' (1988) 86 Michigan Law Review 827–91.
R Posner, Economic analysis of law (7th ed, Aspen 2007).
CJ Mayer, 'Personalizing the impersonal: Corporations and the bill of rights (1990) 41 Hastings Law Journal 577–650.
D Fagundes, 'Note, What we talk about when we talk about persons: The language of a legal fiction’ (2001) 114 Harvard Law Review 1745, 1766.
D Millon, ‘Theories of the Corporation’ (1990) Duke Law Journal 201-262.
RS Avi-Yonah, The Cyclical Transformation of the Corporate Form: A Historical Perspective on Corporate Social Responsibility (2005) 30 DEL J CORP L 767, 767.
R Harris,’The Transplantation of the Legal Discourse on Corporate Personality Theories: From German Codification to British Political Pluralism and American Big Business’ (2006) 63 WASH & LEE L REV 1421, 1424.
MJ Phillips, ‘Reappraising the Real Entity Theory of the Corporation’ (1994) 21 FLA ST U L REV 1061, 1067–68.
D Millon, ‘Theories of the Corporation’ (1990) Duke Law Journal 201-262.
J Armour, H Hansmann, R Kraakman, ‘The Essential Elements of Corporate Law: What is Corporate Law?’ in Reinier Kraakman and John Armour and Paul Davies and other (eds), The Anatomy of Corporate Law: A Comparative and Functional Approach (2nd edn, 2009) 6.
FH Lawson, ‘The Creative Use of Legal Concepts’ (1975) 32 New York University Law Review 907, 915.
HG Henn and JR Alexander, Laws of Corporations and other Business Enterprises (3rd edn, 1983) 144-45.
JC Coates, 'State Takeover Statues and Corporate Theory: The Revival of an Old Debate' (1989) 64 NYU L REV 806, 810.