Legal Personhood for AI systems: Should they ever hold limited legal status, as companies, rivers etc. do?
Geethika Sreekumar · Published 2026-11-20T18:45:33.226Z · Version 1
Author profile: Geethika Sreekumar
Abstract
The fast scale development of various Artificial Intelligence (AI), specifically the autonomous and generative systems challenge the traditional understanding of what we mean by legal personhood. Traditionally, legal personality has been associated with human beings and artificial entities such as corporations. But religious icons and, in some jurisdictions, even rivers and ecosystems, have also been granted legal personhood. These developments raise an important question: ‘if the law can recognize non-human entities as legal persons, should artificial intelligence systems be similarly recognized?’. This paper examines legal personality through corporate, religious and environmental jurisprudence and then critically evaluates the arguments for and against AI personhood. It considers the 2017 proposal of the European Parliament for electronic persons, the resulting risk-based regulatory framework of the European Union, and the announcement of symbolic citizenship for Sophia in the Kingdom of Saudi Arabia. Real world situations like autonomous vehicles and generative AI which is very prevalent nowadays clearly depict these challenges of assigning accountability and responsibility across developers, manufacturers and its end users. This paper will try to draw conclusions focusing on India and go on arguing that an independent personality for an AI is currently not necessary and might be in fact dangerous. It could also create a gap in accountability by completely shifting responsibilities to an artificial system without any meaningful asset or capacity to address the harm caused. Instead, this paper will suggest a regulatory approach for high-risk AI systems which is underpinned by registration, insurance, audits and traceability along with human oversight. Finally, the paper concludes that technological autonomy does not infact justify legal independence of AI. The main challenge here is to hold the actors behind these AI systems accountable.
INTRODUCTION: When machines challenge the meaning of ‘Legal Personality’
Artificial intelligence can no longer be seen as limited to repetitive calculations or pre-defined commands. Most modern systems can write legal documents, analyze various financial markets, recommend medical treatments, interact with consumers, and even control complex machines. Few of the systems also has the ability to select and execute actions involving minimum human involvement. This development is problematic for legal systems that are built merely on human decisionmakers and identifiable organizations.
The question here of ‘whether AI ought to be a legal person’ is often posed by analogy. Corporations are legal persons[note]; religious idols have been recognized as juristic persons in India[note]; and the Whanganui River has been recognized as a legal person in New Zealand[note]. If legal personality is not confined to biological life, the argument goes, then clearly autonomous AI could also be eligible for this classification.
That analogy is not quite right. Legal personality is not a reward for intelligence, autonomy or human-like behaviour. It is a legal technique which is used to achieve certain purposes. The corporate personhood allows for organized economic activity, the personhood of a religious idol can protect endowed property, the personhood of a river can provide representation for ecological interests and in turn help in protecting the environment. The pertinent question, therefore, is not whether AI is like these entities, but whether giving personality to AI would offer a better solution to a real legal problem than the existing legal mechanisms.
The following analysis suggests that AI systems are not currently in a position to receive legal personhood. While their independence may make it harder to demonstrate causation and liability, these systems are created, provided, placed into use, serviced, and used for commercial gain by human and organizational actors. If we grant such entities legal personhood, the actors can be allowed to pass on liability to an entity that has no assets of its own and does not even comprehend what legal duties mean. A compromise is needed in this case.
What does Legal Personality actually mean?
Legal personality[note] entails the recognition of an entity as one that can enjoy various rights, duties, powers, and liabilities in the eyes of law. It must not be confused with legal capacity or agency. Personality involves whether an entity is recognized as a subject of law, while capacity involves what rights and obligations can it enjoy or exercise. On the other hand, agency involves the capability to act on behalf of another individual. An agent is capable of doing legally important acts without having personal legal status. The distinction is necessary since the use of AI in legally relevant activities has been ongoing.
Section 10A[note] of India’s Information Technology Act 2000[note] bars a contract from being invalidated on the basis of having used electronic means in establishing the contract. Furthermore, Section 11[note] sets out the rules that determine attribution of electronic records, including those automatically produced by an information system.
Independent decision-making also does not necessarily make the AI legally responsible. The choice made by the system does not necessarily have to understand the legal ramifications of it. In the case where an AI device provides an inappropriate plan or outcome that is illegal, the legal question that needs to be answered is the identity of the designer of the software, the knowledge of risk, the selector, warning, etc. Calling the AI responsible there will not solve the issue that has raised.
Lessons from existing non-human but legal persons
3.1 Personality as an organizational took: ‘Corporations’
The basis of the principle is Saloman v. A Saloman & Co Ltd[note], where the House of Lords ruled that the incorporated company has an independent legal personality of its own distinct from the personalities of its shareholders. The company may own property, contract, sue and be sued in its name. Similarly, section 9[note] of the companies act 2013[note] in India attributes a body corporate status to the incorporated company. A corporate personality is backed up by such institutional elements as incorporation, governance, assets, management, statutory obligations and means of enforcement. It regulates the activities of humans in the economic sphere and allocates the risks.
An AI algorithm itself is not usually an organization with shareholders, management and capital. And in case of losses caused by an independent trading system, the first questions are who provides it with the capital, who governs its assets, who acts as its representative and who pays the compensation of a judgment in case of its insolvency? The above comparison clearly shows that the principle of corporate personality is helpful, but it is dependent upon the appropriate legal and financial environment.
Corporate personality in itself cannot protect human beings. Director’s duties, fraud provisions, insolvency law and statutory liability coexist with the principle of separate personality. Therefore, any AI framework must keep the accountability of the system’s owners and operators.
3.2 Personality through representation: ‘Religious entities’
In addition to this, Indian law recognizes some religious institutions as juristic persons. The legal position of a consecrated Hindu idol and the legal position of the human shebait who serves the deity was considered by the Privy Council in Pramatha Nath Mullick v Pradyumna Kumar Mullick[note]. In Shiromani Gurdwara Parbandhak Committee v Som Nath Dass[note], the Supreme Court recognized Guru Granth Sahib as a juristic person in the appropriate religious context.
Thus, these judgments demonstrate that being alive and capable of communicating is not necessarily a requirement of legal personality. The idea is not to recognize the independent decision-making. The recognition of legal personality helps in protecting and administering religious rights and properties with the help of human representatives. The analogy raises a question for AI regarding whom the representative will represent and in whose interest. If the representative is controlled by the developer or owner, then independence will be questionable.
3.3 Personality as a protective mechanism: ‘Nature and rivers’
New Zealand’s Te Awa Tupua (Whanganui River Claims Settlement) Act 2017[note] confers legal personality on the Whanganui River, entailing the rights, powers, duties and liabilities of a legal person. In this case, Te Pou Tupua will be the human voice and face of the river. In India, the Uttarakhand High Court held in Mohd Salim v State of Uttarakhand[note] that the Ganga and Yamuna rivers and some associated entities are legal persons. This decision was subsequently stayed by the Supreme Court in July 2017[note]. It is thus misleading to call the status of legal personality conferred by the High Court judgment as an operative status in India. The above comparison shows that there is a distinction in the purposes served by the legal personality of rivers and artificial intelligence (AI). The problem of representation of interests serves as the rationale behind the grant of legal personality in the former case, whereas the latter arises out of the creation of such persons by organizations which have business interests. These interests need not be unrepresented. The grant of personality in such cases will only protect the organizational interests at the cost of the individuals who are harmed by the use of AI.
International Developments: From electric-personhood to a risk-based regulation and approach
4.1 2017 resolution of the European Parliament
Paragraph 59(f)[note] of the European Parliament’s Resolution on Civil Law Rules in Robotics of 2017[note] considered the possible creation of a particular legal status for autonomous robots in the long run, which may involve electronic personhood when robots operate autonomously or communicate with third parties. However, this was just a resolution and not a piece of EU legislation introducing electronic persons. This raised fears that AI could be utilized by manufacturers/operators to evade responsibility.
4.2 Product liability reform and the EU AI act
Subsequent regulatory trends in the EU illustrate this point well. Regulation (EU) 2024/1689[note], Artificial Intelligence Act[note], establishes a risk-based regime and imposes requirements on those providers, deployers and other operators who are responsible for such systems. Depending on the particular system and functions, the regime covers risk management, documentation, transparency, human supervision and monitoring. It regulates individuals and companies who are responsible for the placing of these systems on the market and for their deployment rather than creating an independent category of artificial intelligence persons.
Directive (EU) 2024/2853[note] on product liability for defective products explicitly incorporates software, including AI systems, into the product liability scheme. Relevant software providers are viewed as manufacturers under this Directive. It applies to products which were placed on the market or brought into use after 9 December 2026.
4.3 Sophia: The symbolic citizenship
According to reports from 2017[note], Sophia, an artificial intelligence machine created by Hanson Robotics, was declared a citizen of Saudi Arabia. This made international headlines, but it did not set up a new legal framework for what it means to be a citizen of AI. Being a citizen and being a person under law are two separate things. It is a statement, but it does not say whether the robot has the ability to vote, own property, receive an inheritance, be punished, or have responsibility.
Practical accountability problems
5.1 Autonomous vehicles and physical injury
The case of 2018 where an Uber autonomous car hit and killed Elaine Herzberg in Tempe, Arizona[note] is an example of how hard it is to allocate liability for accidents in autonomous transport systems. In this case, the victim was hit by the car while its driving mode was autonomous but at the same time, a human safety operator was on board. Failures were found in both automated driving system and safety oversight in the investigation by the US National Transportation Safety Board. The case clearly shows that the problem is not just about the responsibility for the actual moment of driving but also the way the vehicle was designed, tested, operated, supervised and the safety culture of the company. In case of accident caused by a completely autonomous vehicle, granting it with the status of a legal person will help find a nominal defendant but will not solve the problem of compensation.
5.2 Generative AI and inaccurate, misguiding information
In Moffatt v Air Canada[note], a chatbot provided a customer with incorrect information about bereavement fares. Air Canada argued, among other things, that the chatbot was a separate entity responsible for its own actions. The British Columbia Civil Resolution Tribunal rejected that position and held the airline responsible for the information provided via its chatbot. The decision is useful because it illustrates that An organisation cannot simply escape responsibility by describing its chatbot as being separate from itself.
The same problem can arise where generative AI produces an incorrect legal deadline, a defective contract clause or a misleading financial explanation. Liability may depend on representations made by the provider, the role of a professional or deployer, the user’s reliance and the applicable contractual or consumer-protection rules. Treating the AI as the primary defendant would not identify the human actor who caused the output to be produced, who prompted it, who supplied it, who deployed it or who distributed it.
5.3 Algorithmic Biases
AI systems those are trained with historical data could further perpetuate or even amplify the inequalities that exist. A recruitment model trained on previous hiring decisions might disadvantage candidates from groups that have traditionally been excluded from some roles[note]. A credit-scoring model may include variables that are proxies for socio-economic disadvantage. In 2016, ProPublica looked at the COMPAS recidivism-risk tool and found that it raised concerns about racial disparities in risk classification[note].
The problem is not that an algorithm might be wrong. The output might appear to be neutral and objective but it also encodes choices about data, acceptable errors etc… Personhood for AI could thus interpose another actor or entity in between the individual who is affected and the institution that had selected and relied upon that system.
5.4 Reputational harm, deepfakes and impersonation
Generative systems can produce realistic audio, images and video of people doing or saying things that never occurred. Deepfakes can be used for fraud, defamation, political manipulation and non-consensual intimate imagery. Responsibility could fall on the creator or distributor of the content itself, the platform that allowed the content to circulate, or the service provider that did not meet the relevant obligations. The legal analysis must differentiate between the model, the user, the hosting platform and the distributor. By treating the model as a defendant, these differences are obscured[note].
Arguments for AI Personhood and its limits
6.1 Autonomy and unpredictable conduct
The most compelling argument for AI personhood is that the increasingly autonomous systems could act in ways that are hard to predict and not attributable to a specific human instruction. A separate legal identity could, theoretically, provide a direct defendant. But the difficulty of proving causation or fault does not mean that there is no responsible legal actor. For negligence, product liability, statutory duties and contractual rules to apply, the exact harmful result does not have to be intended by a human. Responsibility can be spread out over the lifecycle.
6.2 Assets and compensation
A sophisticated AI system could be used to generate revenues or manage assets, which could mean insurance can be maintained and victims compensated. But personality does not provide sufficient resources. There may be cases where a company might put very little assets into an AI system which leaves it unable to meet the claims. Issues like insurance, guarantees and various capital requirements might directly address the compensation part alone considering no other alternatives.
6.3 Possible machine consciousness
A tougher argument is about the possibility of future artificial intelligence systems having subjective experiences or morally relevant interests. AI systems quickly grasp human language, emotions or even consciousness through conversations and various advanced systems have already proved to appear human-like even though there is no direct evidence of sentience in themselves. But if any arguable scientific evidence emerges that even AI can have these humane interests, the law my then have to quickly rethink whether the exisiting protective rights are justified and enough. Even then, the limited protections will not automatically give them any legal personality.
When, why and at what stage might this concept of AI personhood turn dangerous?
7.1 The shifting risk of accountability
The main risk here is that an AI personhood might turn as a liability shield instead of its actual purpose. If for example, a company profits from a particular system, then why not argue that this system acted independently and therefore are separately responsible? This is not inherent to personhood. The legislation can preserve the liability for the operators and developers of the systems. But, if the same goals in practice can be achieved without a need for separate category of legal person, then the risk creation will accordingly shift.
7.2 Intention, culpability and punishment
Criminal law requires proof of certain mental elements such as intention, knowledge or recklessness. As for AI systems, these may produce harmful outputs. But is the intention or culpability in the sense humans understand it, attributable to the AI systems? This remains to be determined. What about software? What does it mean to imprison software? Would its elimination (the death sentence for software) be the same as imprisonment? These questions illustrate why the responsibility for unlawful acts of AI systems cannot simply be imputed to machines. Unlawful human actors who design, deploy and/or abuse AI systems must remain accountable under the law.
7.3 Regulatory arbitrage and divided responsibility
AI systems’ supply chains encompass jurisdictions. Developers, servers, operators, and victims can be in different jurisdictions. Adding layers to the AI system means separating entities, creating opportunities to place the nominal defendant in a jurisdiction with weaker enforcement. The law should be focused on entities that create and introduce systems into the market. They should retain obligations that are enforceable.
7.4 Misplaced trust and anthropomorphism[note]
Conversational systems can be designed to appear empathetic, confident and independent. Granting them legal recognition will likely strengthen the misconception that systems possess true human-like understanding or that their outputs are accurate. Legal frameworks should encourage recognition of the limits of a system and not induce the public to think that fluent simulation stands in for human judgment.
India: Existing legal foundations and regulatory gaps
8.1 Companies and juristic persons
Incorporation, according to Section 9[note] of the Companies Act 2013, endows a company with its own legal personality, and the ability to own property, and the power to initiate and face legal proceedings. Though cases involving the legal personality of deities, among others, demonstrate that legal personality can exist beyond the domain of incorporated companies, this fact doesn’t automatically bestow the same legal status on artificial intelligence. Artificial Intelligence systems don't operate like religious entities. The legal system for one has its foundations in something quite different.
8.2 Information technology Act, 2000
Section 10A[note] and 11[note] of the Information Technology Act 2000 lay the ground rules for the formation of contracts electronically and for attributing electronic messages, such as those sent by automated systems, to their originators. Although useful in the context of AI transactions, the Sections fail to provide for the loss of a physical nature or any other form of loss resulting from autonomous systems and AI. The Sections also fail to cover algorithmic bias or products based on defective AI.
8.3 Data Protection and criminal law
Data fiduciaries are subject to obligations regarding digital personal data under the Digital Personal Data Protection Act of 2023[note]. The Act is not a comprehensive AI liability statute, which is significant because AI systems frequently depend on personal data. It does not, by itself, address every problem pertaining to discriminatory automated decisions, software flaws, or physical harm. The Bharatiya Nyaya Sanhita 2023[note] defines "person" as a company, association, or group of people, whether or not they are incorporated, according to Section 2(26)[note]. It does not resolve whether an AI can have the mental components necessary for a specific offense, nor does it explicitly establish AI systems as independent legal persons.
As a result, India lacks a comprehensive, cross-sector framework that addresses the entire lifecycle of high-impact AI systems, but it does have adaptable legal tools. Targeted regulatory development is a better course of action than declaring electronic personhood right away.
A proposed alternative: Registered autonomous system status (RASS)
In order to give certain high-risk AI systems a formal regulatory identity without giving them independent legal personality, this paper suggests a framework known as Registered Autonomous System Status (RASS)[note]. This system will intend to serve as a base mechanism for the purpose of registration, traceability etc… rather than considering AI as a separate entity with a different set of rights and liabilities. High-risk AI systems should be registered with the relevant regulatory body prior to deployment, providing information about their developers, manufacturers, deployers, intended uses, approved uses, model versions, and accountable organizational representatives. To allow regulators, courts, and impacted parties to track the system's operation and identify the parties involved, each system should be assigned a unique identification number. The AI system would not become a legal person through such registration, which would be comparable to the regulatory identification of cars or medical equipment.
Mandatory insurance and financial guaranties commensurate with the system's risks would be a crucial component of this framework. For example, third-party liability insurance may be necessary for autonomous transportation systems; professional or product liability coverage may be necessary for AI applications used in healthcare; and financial guaranties may be required for systems handling significant financial transactions. This would ensure that the people affected by these AI systems now have a possible path to compensation without having to face and settle the complex disagreements about true accountability. Liability should also be distributed by taking into account each actor's level of control, their awareness of possible risks, the financial gains they make from the system, and their capacity to stop or lessen the harm.
AI to be developed along with innovation
Every stage of an AI’s lifecycle must be subject to some sort of regulation. The providers should evaluate thoroughly the effects, cybersecurity concerns etc… at the development stage itself. High-risk systems should be tested in environments similar to their intended use before being deployed. Deployers should keep an eye on performance, incidents, and risk shifts while in use. Reassessment should be prompted by significant updates that significantly change risk or capabilities. Systems that pose unacceptable risks ought to be able to be withdrawn, suspended, or restricted.
Innovative AI applications in healthcare, finance, education, and transportation should be tested under supervision in regulatory sandboxes in India[note]. Sandboxes shouldn't be exempt from laws pertaining to consumer protection, safety, or fundamental rights. They ought to have clear reporting guidelines, compensation plans, and limits.
Additionally, cross-border collaboration is crucial. India should look into ways to look into AI-related incidents, obtain technical evidence, identify suitable safety standards, and stop regulatory arbitrage. The end goal must be to achieve a minimum accountability standard that continue to be effective even when the AI cross borders.
The question still remains: what if AI become conscious?
Even if now the idea of legal personhood for AI is argued to be rejected, it doesn’t mean that it is a forever rejection. The ethical and legal debate will continue and shift towards finding trustable scientific evidences which might prove that an AI system can have or had experiences similar to the other legal entities. It would still be very much necessary to draw a distinction between three important questions:
(i) Whether the AI system has true morally significant experiences?
(ii) If yes, whether those experiences warrant legal protection?
(iii) And, whether the legal personality is the most effective way of doing so?
Conclusion: Accountability, not artificial personhood
Legal personality constitutes a very flexible legal construct rather than a status which is exclusive to biological humans, as evidenced by its application to corporations, religious organizations, and rivers through various case laws. However, legal personality is mostly conferred to fulfil specific functions. For example, river personhood represents ecological interests, corporate personality facilitates organized commerce, and religious juristic personality protects the religious property and interests.
AI presents a different challenge. Human and institutional actors are responsible for designing, supplying, deploying, and profiting from AI systems. Although increasing autonomy might complicate the determination of causality and fault, it does not mean that technology operates independently of other human institutions. Granting independent legal personality to AI could thus result in uncertainty and eventually lead to the evasion of accountability. Therefore, at this point, India should refrain from giving AI systems any such independent legal personality or personhood.
Instead, it should try and create a reasonable regulatory framework based on insurance, financial guarantees, registration, traceability, audits etc… to prevent AI caused harms from arising. The proposed Registered Autonomous System Status offers a workable compromise, acknowledging the regulatory importance of autonomous systems without treating them as a separate legal entity. If any new evidence shows that artificial systems have morally significant interests, the law must be re-examined at that later point.
However, the current objective shouldn't be distracted by that possibility. Legal independence is not merely a result of technological autonomy. Even if a machine or AI makes a decision without any human involvement, the decision would still be part of a chain. Therefore, the main issue is still not whether AI should be considered a legal person in all sense, but it is instead whether the existing laws can guarantee that no person or entity can avoid accountability by merely putting an intelligent machine between itself and the results of its own actions.
Footnotes
- Saloman v Saloman Co Ltd (1897) AC 22
- Pramatha Nath Mullick v. Pradyumna Kumar Mullick (1925) 27 Bom LR 1064
- Te Awa Tupua (Whanganui River Claims Settlement) Act 2017
- ‘ Legal personality’ ( Merriam-Webster Legal Dictionary)
- Information Technology Act 2000, s 10A
- Information Technology Act 2000
- Information Technology Act 2000, s 11
- Ibid (n 1)
- The Companies Act 2013, s 9
- The Companies Act 2013
- Ibid (n 2)
- Shiromani Gurdwara Prabandhak Committee v Som Nath Dass (2000) 4 SCC 146
- Ibid (n 3)
- Mohd Salim v State of Uttarakhand 2017 SCC OnLine Utt 367
- State of Uttarakhand v Mohd Salim SLP (C) No 016879/2017 (SC, 7 July 2017)
- European Parliament’s Resolution on Civil Law Rules in Robotics of 2017, Para 59(f)
- European Parliament’s Resolution on Civil Law Rules in Robotics of 2017
- Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024, EU AI Act
- Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024
- Directive (EU) 2024/2853
- United Nations, “ At UN: Robot Sophia joins meeting on artificial intelligence”
- “ Uber’s self-driving operator charged over fatal crash (BBC News, 16 th September 2020)
- Moffatt v Air Canada 2024 BCCRT 149
- Dr Natalie Sheard, “Discrimination by recruitment algorithm is a real problem”, 15 May, 2025
- Jeff Larson, Surya Mattu, Lauren Kirchner and Julia Angwin, “How we analysed the COMPAS recidivism algorithm”, (May 23, 2016)
- Farid H, “Mitigating the risks of manipulated media: Confronting deepfakes and digital deception” (2025) 4 (7) PNAS Nexus
- D Motas-Rojas and Others, ‘Anthropomorphism and its adverse effects on the distress and welfare of companion animals’ (2021) 11(11) animals 2363
- Ibid (n 9)
- Ibid (n 5)
- Ibid (n 7)
- Digital Personal Data Protection Act 2023
- Bharatiya Nya Sanhita 2023
- Bharatiya Nyaya Sanhita, 2023, 45 of 2023, (India), 2023, s 2(26) <https://prsindia.org/files/bills_acts/bills_parliament/2023/Bharatiya_Nyaya_Sanhita,_2023.pdf> accessed 2026-11-19.
- Autonomous system numbers https://www.iana.org/assignments/as-numbers
- RBI, Inter-operable regulatory sandboxes, https://www.rbi.org.in/commonman/English/scripts/FAQs.aspx?Id=3822
Bibliography
Bharatiya Nyaya Sanhita, 2023, 45 of 2023, (India), 2023, 63, Exception 2 <https://prsindia.org/files/bills_acts/bills_parliament/2023/Bharatiya_Nyaya_Sanhita,_2023.pdf> accessed 2026-11-19.
Saloman v Saloman Co Ltd (1897) AC 22
Pramatha Nath Mullick v. Pradyumna Kumar Mullick (1925) 27 Bom LR 1064
Te Awa Tupua (Whanganui River Claims Settlement) Act 2017
‘Legal personality’ (Merriam-Webster Legal Dictionary)
Information Technology Act 2000,
The Companies Act 2013
Shiromani Gurdwara Prabandhak Committee v Som Nath Dass (2000) 4 SCC 146
RBI, Inter-operable regulatory sandboxes, https://www.rbi.org.in/commonman/English/scripts/FAQs.aspx?Id=3822
Digital Personal Data Protection Act 2023
Bharatiya Nya Sanhita 2023
Autonomous system numbers https://www.iana.org/assignments/as-numbers
D Motas-Rojas and Others, ‘Anthropomorphism and its adverse effects on the distress and welfare of companion animals’ (2021) 11(11) animals 2363
Farid H, “Mitigating the risks of manipulated media: Confronting deepfakes and digital deception” (2025) 4 (7) PNAS Nexus
Moffatt v Air Canada 2024 BCCRT 149
Dr Natalie Sheard, “Discrimination by recruitment algorithm is a real problem”, 15 May, 2025
Jeff Larson, Surya Mattu, Lauren Kirchner and Julia Angwin, “How we analysed the COMPAS recidivism algorithm”, (May 23, 2016)
Mohd Salim v State of Uttarakhand 2017 SCC OnLine Utt 367
State of Uttarakhand v Mohd Salim SLP (C) No 016879/2017 (SC, 7 July 2017)
European Parliament’s Resolution on Civil Law Rules in Robotics of 2017
Regulation (EU) 2024/1689 of the European Parliament and of the Council of 13 June 2024, EU AI Act
Directive (EU) 2024/2853
United Nations, “At UN: Robot Sophia joins meeting on artificial intelligence”
“Uber’s self-driving operator charged over fatal crash (BBC News, 16th September 2020)
Research declarations
Funding and sponsor involvement
No external funding
Competing interests
None declared
Author contributions
Geethika Sreekumar- Main contributor of research and writing
Originality and related publications
None
Methods and legal currency
Doctrinal and comparative method chosen primarily relying on case laws.
Limitations
Indian jurisdiction referred to with slight comparison made to the European Union directives
Ethics and participant protections
None
Data and materials availability
-
AI and writing assistance
AI was used for grammar, improving language and refining structure. The research, content, arguments etc... remain the responsibility of the author.
Article citation
Geethika Sreekumar, “Legal Personhood for AI systems: Should they ever hold limited legal status, as companies, rivers etc. do?”, The Legal Gazette — Open Scholarship (2026-11-21), version 1, https://www.thelegalgazette.org/scholarship/2bb747e1-5050-4e72-b9ce-10d9dbb5c11e.
Rights & licence
Copyright owner: Geethika Sreekumar
Selected licence: CC-BY-NC-4.0
Access and reuse
Free to read — noncommercial reuse licence
Commercial reuse is restricted. This is not unrestricted reuse and does not meet TLG’s preferred CC BY licensing target.
