R. ARUN KUMAR
Department of Labour and Administrative Law
Chennai Govt. Law College, Pattaraiperumpudhur, Chennai
Change...Change is the only constant. All that changes lives and all that doesn’t, perishes. The developmental changes across the globe are inevitable. Neither the law nor the society has the ability to stop them; rather they must try to cope up with the dynamicity. If either of the two fails, the other comes to forcibly rescue the former. The concept of technology is no exception to that. The first thing to be discussed is the inspiration for the title of the paper. The phrase “4th Industrial phase” indicates the era of Artificial Intelligence (hereinafter referred as AI), the word e-personhood has two sources, firstly, it was first coined in 1967 in the Life Magazine[i]and secondly, the concept became popular when it was used in the Draft report submitted to the European Parliamentary Committee on Legal Affairs[ii]. The term agent-hood relates to the contractual regime. The author would like to strictly focus on the legal status and aspects of AI without going deep into the technical and philosophical arena of AI.
It is important to get clarified on the difference between Robots and AIs. The former are just machines which are assigned with a specific task and the latter is the intelligence exhibited by those machines, which mimic human and carry out tasks with human cognitive skills. So the striking difference is the presence of human trait in AI. The academic definition[iii] which is widely accepted is “It is the science and engineering of making intelligent machines, especially intelligent computer programs. It is related to the similar task of using computers to understand human intelligence, but AI does not have to confine itself to methods that are biologically observable”.[iv]
There are 4 types of AI, classified based on the ability to do tasks, namely, Reactive machines, Limited memory, Theory of Mind (ToM) and Self-awareness machines. Reactive machines don’t have memory, they can't recall the past and use it for the current task. Every task is a new task for it. They react based on responses, e.g. IBM’s DEEP BLUE chess game. Limited memory, as its name suggests, is programmed with limited memory and use it for the tasks,e.g. self-driven cars. ToM is a humanoid robot with larger memory but programmed. It is not a practical but theoretical robot, because it doesn’t understand humans. It has decision making power but is not autonomous,e.g. MITRA, ROSS. Self-awareness machines are humanoid plus practical robots that can understand the feelings of human; they are autonomous and have decision making power, e.g. Sophia. As earlier said the paper focuses on the legal aspect of the AI, it revolves around 3 factors based on which each concept in this paper will be discussed. There are morality, social reality and Legal convenience.[v]
CONCEPT OF E-PERSONHOOD:
It is well known that the status of a natural person to AI is impossible due to biological reasons and the only status of a legal person is possible. The trend of granting legal personhood relates backs to the 13th century where the Pope Innocent proposes the Doctrine of persona fit allowing monasteries to have legal existence without including the monks.[vi]This gradually developed during the course of time to include Sovereign states, International organizations like UN, EU[vii], Companies, Hindu idols[viii], WhangareiRiver[ix] within the purview of the legal person. Likewise, AI designers and legal scholars strive hard to recognise AI as a legal person. The status of legal person does not depend on the natural person, but, as already mentioned, on the 3 factors.
Under the first factor, two question arises, what realities are entitled to legal personhood? And what are characteristics need to be possed by such realities? The answer to the second solves the first. Companies, associations, boards are all examples of legal persons, the common thing they possess is that they can act autonomously, make decisions, are self-aware of their existence and have subject experience about the world. If AI robots possess these traits then it can be considered to be eligible for legal status. Among the 4 types, only the limited memory, ToM and self-awareness machines are capable of decision making. On seeing the criteria of subject experience, only the later passes the test, leaving behind the former two, which lacks sentience and their autonomous act is just a programming.
The second factor is societal constrains, recently bots have been frequently used for both commercial and non- commercial transactions without which the market and the economy slows down. If this continues, during the course of time, people might recognise such bots as actually engaging themselves in the transactions. This would lead to a scenario where the society or the market transactions to be completely dependent on the bots like AI. Thereby pressuring the law to recognise them.
The third factor is the legal convenience, the reason behind granting ships, associations the status of a legal person is that they can be kept under the surveillance or supervision of law, will also be easier in liable issues and thereby serving a legal purpose at cost-efficient.[x] As a conclusion for this part, the reactive machine passes none and self-awareness machine passes all the criteria. While the other two AI failed to satisfy the morality factor.
ACKNOWLEDGEMENT OF RIGHTS:
Next legal aspect which needs to be discussed is a bundle of rights and duties comprising the right to property, right to sue and be sued. They are given to Companies, trusts, idols, etc because it services the purpose for which they exist. If taking property rights as an example, why does an AI require property? Does a non-human can morally hold property? Though a firm is non-human, it is an association of humans. Since property includes IPR, there was an instance where a monkey named Naruto captured some photos from a still camera. Whether copyright laws[xi] applicable to these photographs? Whether a non-human just by clicking a button makes him an artist or author of its work? High tensions prevailed in U. S. Courts between the U. S. copyright office[xii] and the PETA. The court finally negatived the contentions of the PETA.[xiii]
On considering the social impact, the criteria to hold a property is the presence of subject experience over the property, i.e., if that monkey can feel the existence of a property and reason for that existence, then it can be the owner of that property. With that said, the reactive machines and limited memory are out of the contest, though ToM has some experience but it is due to the programming by the designer, and leaving behind the self-awareness machine that has a sense of realisation of its existence, its property and the world.
It makes the legal area difficult to grant the property rights to AI because a lot of changes must be brought to the existing legal framework like property laws, IPR laws, etc. It also poses challenges to taxation, agent-principle relation under the contract laws and may allow the designers or users to misuse the AI’s properties like the Benami transactions. From this, we can derive that all the three factors did not support the granting of the property right to AIs, except the social realities, lean towards the self-awareness AI.
CITIZENSHIP AND GENDER:
It is ironical to assume that an AI robot is superior than the women. But it is the reality happened in Saudi Arabia recently in 2017, where Sophia, asocial humanoid robot which was granted citizenship and gender. This erupted a lot of controversies among the legal scholars and technicians, saying it was against nature. The author opines that the grant of citizenship was timely done to encourage the 1st of its kind as Self-Awareness AI, which can express feelings, crack humour with distinct facial expressions, follow faces, sustain eye contact, recognize individuals, able to process speech and have conversations using a natural language subsystem[xiv]. This must be considered as an encouragement step for future developments. On the other side, there are many discouragements in granting such a status to robots. First thing is that the AI lacks human-like identities like iris, fingerprints, DNA, etc. It can be said that it possesses unique bar code, skin mask, audio mark in her voice and also electromagnetic signals similar to that of human brain waves. These are all technological identity belonging to the hardware and not Sophia’s identity.[xv]
With the grant of citizenship status, the second issue is, whether political rights follows AI? Whether he or she can vote in a poll? Can she or he participate as a candidate in an election? Going to an extreme, by winning the election can he or she rule a nation? Before answering these, there arises another problem as to who would vote, AI or the user? If it is autonomous it can vote, but if the user inputs any programme into AI regarding as to whom the vote should be cast, it would be a disaster. These are the few opposing points that are raised against the grant of citizenship.
Following the grant of citizenship there follows the right to gender determination. For this, it is important to know the difference between gender determination and sexual orientation, the former means one’s own perception on his or her own sex, in other words, what a person thinks of him about his own sex. While the latter deals with one’s own wish or choice to choose his life partner.[xvi] If the law doesn’t have an issue in granting the citizenship then it should be automatically followed by the right to gender. In the case of Saudi Arabia, where the nation denies women certain rights, recognizing a robot as a female citizen but not denying those rights depicts the natural female citizens inferior in society even below the AI. The government of Saudi Arabia justifies it by saying that the women rights will be gradually improved under the VISION 2030[xvii]. This makes the women more pitiful in Saudi Arabia.
Some communities make a controversial view that AIs are biologically and naturally incapable to marry and mate with a partner. If bodily capability or hormone presence is the sole criteria, then an impotent person is neither a male nor a female, according to those who hold that view. So gender determination is not based on the outside, but comes from inside. That’s why one needs clarity between sex and gender.
There are two assumptions that can be derived regarding the liability for the acts of AI. One is either the AI or the designer/user is liable but not both, another is, one bears responsibility only when it has an autonomous state and sense of sentience. So legally minors, unsound persons can’t be made liable. Likewise, whether the AI be granted exemption simply because they are created by humans and making the latter liable for the acts of former.
We can blindly say the reactive machine lacks the autonomous state, therefore, the liability is with the user or the designer. On the other extreme, there is self-awareness machine with satisfactory sentience coupled with evolving conduct, so the liability shifts to the AI itself for its own autonomous act, except defect in coding or programming.
The remaining two, i.e., the ToM and limited machine, the onus of liability is quite complex. Both possess autonomous status but are of two kinds, defect in program or code and AIs evolving conduct. For the first one, we can compare AI to a pet, if it causes a nuisance to others, the owner is held liable because the pet doesn’t have an autonomous mind. Likewise, the AI of the above two types though possess autonomous state; they are disrupted by the defect in coding, which makes the owner/user liable. Under no circumstances, the owner can be excused from liability saying AI has an autonomous mind, where he himself had made a defect in programming.
In the second type where the AI has evolving conduct, it is solely liable for its acts. The designer while manufacturing the AI robot or program does all his best with due care and caution to prevent all probable risk arising out it and with good faith brings to the market[xviii]. Here he can’t be made liable for the acts of AI without any faults on his side. If he is made so, the designer will undergo a fear towards his invention and thereby stalls his technical progress. So AI is solely liable here[xix].
If we take the self-driven car as an example, there is no fault with the designer but the car makes an accident causing harm to the victim. In this case who will pay damages to the victim? If the designer is exempted from liability how will the AI compensate the victim? This is a clash of interest between the designer and the victim.
To solve this problem the concept of Insurance scheme must be applied where the AI must be registered along with the user and designer. The designer must subscribe an insurance policy on behalf of the AI thereby limiting his liability. If no insurance, the designer is borne to full liability. There is another view that only designer must be held liable and not the user, if it is done, the designers would include these future liability expenses in the selling price itself. Ultimately, the burden is on the user indirectly.
Even if the limited liability applied on the AI, where does the property come from? For this the property rights must be offered to the AI, in turn for that right, it must be a citizen of a country and ultimately it requires a personhood status, leaving the rights and status in a cycle.
Apart from compensation for damages, what if AI is liable to penal consequences, does the AI will be jailed or will it be dismantled? If done so, will the AI undergo the true purpose of punishment or the reformation process? Or will it understand the concept of theories of punishment? So from this, we can infer that AI requires a special type of punishment laws.
Earth is not only for humans, but also for nature and future earthling. The latter includes AI. It’s not fair for us to think we are the superior creature surviving in the universe. We should share the earth with others. Views opposing the grant of special status to AI in par with humans arises because of the ego and fear within us. Defect within us should not defer the developmental process, wherein we failed even to recognize whether the Facebook AIs Alice and Bob conversation was a secret one or the malfunctioning. According to the author, granting Sophia, the citizenship right is a revolutionary step and the world should take the aid of it to discover the sciences that are beyond human capacity like 4D and above, parallel universe, life beyond earth, etc. Already AI has been incorporated in sectors like industries (BRABO), Defense (DAKSH), Medicine (DAVINCI, KARMI), banking (LAKSHMI), law (ROSS), etc. On the other hand, the more the autonomous the less the tool it becomes. So it requires all Nationparticipating forum to regulate the field of AI and draft a uniform code for AI like the NPT and also support underdeveloped developing nations via technology transfer. Finally, the answer to the title is that special legal status could be granted to the AIs which is above the legal personhood and below the status of a natural person, provided the nation is ready enough to meet all the future consequences.
1. A. Atabekov, O. Yastrebov, Legal Status of Artificial Intelligence Across Countries: Legislation on the Move, European Research Studies Journal, Volume XXI, Issue 4, 2018, pp. 773 – 782.
2. Alexandre, F. M, The Legal Status of Artificially Intelligent Robots: Personhood, Taxation and Control, Dissertation project, Tilburg University, 12.06.2017.
3. UN World Robotics. Statistics, Market Analysis, Forecasts, Case Studies and Profitability of Robot Investment; Edited by UN Economic Commission for Europe and Co-Authored by the International Federation of Robotics; UN Publication: Geneva, Switzerland, 2005.
4. European Parliament, Resolution (2015/2103(INL)) of 16 Feb 2017.
5. Mathias Risse, Human Rights and Artificial Intelligence An Urgently Needed Agenda, Carr Center for Human Rights Policy, May 2018.
[i]Life was an American magazine published weekly until 1972, as an intermittent "special" until 1978, and as a monthly from 1978 until 2000. During its golden age from 1936 to 1972, Life was a wide-ranging weekly general-interest magazine known for the quality of its photography. [ii] European Parliament's Committee on Legal Affairs, 'Draft Report with Recommendations to the Commission on Civil Law Rules On Robotics' (European Parliament's Committee on Legal Affairs 2016). [iii] Definition by John McCarthy, who founded the artificial intelligence discipline at Stanford and coined the phrase in 1956. [iv]https://www.mitel.com/blog/how-the-experts-define-artificial-intelligence [v]Alexandre, F. M, The Legal Status of Artificially Intelligent Robots: Personhood, Taxation and Control, Dissertation project, Tilburg University, 12.06.2017. [vi]John Dewey, 'The Historical Background Of Corporate Legal Personality' (1926) 35 The Yale Law Journal. [vii]Since the Lisbon Treaty came into force on 1 December 2009. [viii]Pramatha Nath Mullick v. Pradyumna Kumar Mullick Bombay High Court, 27 BOMLR 1064 (Bombay High Court). [ix]Eleanor Ainge Roy, 'New Zealand River Granted Same Legal Rights As Human Being' The Guardian (2017) <https://www.theguardian.com/world/2017/mar/16/new-zealand-river-granted-same-legal-rights-as-human-being> accessed 22 April 2017 [x]Tom Allen and Robin Widdison, 'Can Computers Make Contracts?' (1996) 9 Harvard Journal of Law & Technology. [xi]Hayden Smith, 'Can Monkey Who Took Grinning Self-Portrait Claim Copyright?' Metro (2011) <http://metro.co.uk/2011/07/14/can-monkey-who-took-grinning-self-portrait-claim-copyright-77773> accessed 30 April 2017. [xii] United States Copyright Office, 'Compendium of U.S. Copyright Office Practices: Chapter 300' (2015). [xiii]Naruto, et al. v. David Slater, et al.  United States District Court for the Northern District of California, No 3:2015cv04324 (United States District Court for the Northern District of California). [xiv]https://www.livescience.com/60815-saudi-arabia-citizen-robot.html [xv]https://robohub.org/three-concerns-about-granting-citizenship-to-robot-sophia/ [xvi]NALSA v.Union of India, AIR 2014 SC 1863. [xvii]Saudi Vision 2030 is a strategic framework to reduce Saudi Arabia's dependence on oil, diversify its economy, and develop public service sectors such as health, education, infrastructure, recreation, and tourism. Key goals include reinforcing economic and investment activities, increasing non-oil international trade, and promoting a softer and more secular image of the Kingdom [xviii]Nicolas Petit, 'Law And Regulation Of Artificial Intelligence And Robots: Conceptual Framework And Normative Implications' (2017) <https://papers.ssrn.com/sol3/papers.cfm?abstract_id=2931339> accessed 24 April 2017. [xix] National Science and Technology Council of the Executive Office of the President of the United States of America, 'Preparing For The Future Of Artificial Intelligence' (2016).
We thank you for your support and encouragement to Internationalism. We believe in providing open access to our articles, reports and papers without any paywall for our readers and those who pray well for the think-tank. In order to keep our content open accessed and free, we need your support. Please donate any amount up to 500 INR if possible.
Link to donate: https://pages.razorpay.com/pl_EnnLuU7tqq6lv7/view