INTRODUCTION
The relationship between human civilization and the law has always been dynamic, shaped continuously by technological progress. We are currently witnessing an era where artificial intelligence (AI) has transcended its historical role as a passive computational tool.[1] Modern advancements have given rise to autonomous agents capable of independent decision-making, executing complex contracts, and generating sophisticated digital clones.[2] that mimic human behavior. The proliferation of generative AI models and algorithmic trading systems serves as clear evidence of this paradigm shift.
Consequently, a profound question challenges contemporary legal theory: Should Al agents, clones, and autonomous algorithms be granted distinct legal personhood?[3]
A formal legal analysis must go beyond purely descriptive prose to critically examine whether traditional jurisprudential concepts can accommodate non-human, algorithmic actors. Legal personhood is not merely an ontological recognition of humanity; it is a normative mechanism [4]created by legal systems to assign rights, obligations, liabilities, and legal standing. Historically, law has expanded legal personality beyond natural human beings to include juristic entities[5] such as corporations, idols, ships, and environmental ecosystems.
However, attributing legal rights or liability to autonomous Al agents poses unprecedented challenges to the core tenets of contract, tort, and criminal jurisprudence.z
HISTORICAL BACKGROUND: THE EVOLUTION OF LEGAL PERSONHOOD
To assess whether Al agents or digital clones qualify for legal personhood, one must examine the jurisprudential nature of personality itself. Salmond defines a legal person as “any subject-matter other than a human being to which the law attributes personality.”[6] Legal personality, therefore, is an artificial creation of the legal order rather than a natural fact.
- Theories of Juristic Personality
Under Western jurisprudence, several theories explain how non-human entities acquire legal standing:
- Fiction Theory (Savigny, Salmond): Argues that only human beings possess mind and will. Legal personality granted to non-human entities is a pure legal fiction created by the sovereign state for economic or administrative convenience.[7]
- Realist Theory (Gierke): Asserts that corporate or juristic entities possess a real, collective organism and socio-legal presence independent of state recognition.[8]
- Bracket/Omission Theory (Ihering)[9]: Posits that juristic personhood is merely a bracket placed over a collection of natural human beings to simplify legal relations.
- Concession Theory: Maintains that legal personality flows strictly from statutory authority or state sanction.[10]
When applied to autonomous Al algorithms, the Fiction Theory and Concession Theory offer the most relevant analytical foundations. Since AI lacks organic consciousness, moral agency, and subjective intentionality (mens rea), any legal personality extended to an AI system would necessarily be a legal fiction created by state statutory concession rather than an acknowledgment of innate moral worth.
- Comparative Analogies: Corporate and Environmental Personhood
The corporate entity serves as the primary comparative model for Al personhood. In the landmark case Salomon v A Salomon & Co Ltd,[11] the House of Lords established that a corporation possesses a distinct legal personality separate from its shareholders and directors. Corporations can hold property, enter contracts, and incur liabilities. However, corporate personhood relies on an essential human anchor: directors, officers, and shareholders drive corporate intent and provide capital reserves to satisfy legal judgments. An Al agent lacking financial reserves or human supervision cannot be held accountable under conventional corporate governance principles[12] unless mandatory insurance schemes[13] or digital capital reserves are attached.
Similarly, courts have recognized natural features as juristic entities. In Mohd Salim v State of Uttarakhand,[14] the High Court of Uttarakhand recognized the Rivers Ganga and Yamuna as legal persons, appointing human custodians in loco parentis. In Indian jurisprudence, Hindu idols have long been recognized as juristic persons capable of holding property through human managers (shebaits). While these precedents demonstrate the flexibility of legal fictions, they function primarily to safeguard human heritage or environmental interests through human guardians, rather than bestowing genuine moral autonomy on the non-human entity.
ALGORITHMIC AGENCY, LIABILITY, AND THE DIGITAL CORPORATE VEIL
- Contractual Capacity and Agency Law
Under established classic contract law, a valid contract requires mutual assent (consensus ad idem).[15] In agency law, an agent acts.[16]under the authority and direction of a principal. When an autonomous Al contract engine independently negotiates price and terms, attributing intent (animus contrahendi)[17] directly to the algorithm presents severe legal ambiguity.
Courts historically treated software under the “mechanical tool doctrine.”[18] However, modern deep-learning models operating non-deterministically render the tool doctrine inadequate without updated statutory legal frameworks.
- The “Digital Corporate Veil” and Moral Hazard
A critical danger in granting independent legal personhood to AI agents is the creation of a “Digital Corporate Veil.” Just as corporate status can be misused to shield parent entities from liability, endowing AI clones or autonomous systems with separate legal personhood could allow developers and corporate actors to externalize legal risks.[19] If an autonomous algorithm causes catastrophic financial losses or severe tortious damage,[20] developers could evade liability by arguing that the Al acted independently. Because software algorithms cannot be imprisoned and lack tangible assets, independent Al personhood risks creating a profound moral hazard.[21] and accountability vacuum.
CONTEMPORARY REGULATORY APPROACHES AND COMPARATIVE GOVERNANCE
Global legal orders have begun addressing the Al accountability dilemma through statutory governance rather than extending full legal personhood.
- The European Union Al Act (2024)
The European Parliament and Council adopted the landmark EU Artificial Intelligence Act (Regulation 2024/1689), establishing the world’s first comprehensive risk-based regulatory regime. Notably, the EU explicitly rejected the concept of “electronic personhood”-a proposal previously suggested in its 2017 Civil Law Rules on Robotics report.[22]
Instead of conferring rights or juristic status on Al systems, the EU Al Act classifies Al architectures based on risk categories:
- Unacceptable Risk: Banned applications (e.g., social scoring, cognitive behavioral manipulation). General Purpose Al (GPAI): Subject to systemic risk management and copyright compliance rules responsibility
- The EU framework firmly assigns legal responsibility to providers (developers) and deployers (users), reinforcing the principle that legal liability must remain anchored to human or corporate actors
- High Risk: Subject to strict conformity assessments, continuous risk management, human oversight, transparency, and cybersecurity standards.
- General Purpose AI (GPAI): Subject to systemic risk management and copyright compliance rules.
The Act explicitly places compliance obligations on human providers and deployers.
- The Indian Legal and Regulatory Landscape
India’s regulatory trajectory similarly emphasizes developer/deployer responsibility over Al legal personhood.
- Digital Personal Data Protection Act, 2023 (DPDP Act): Imposes strict obligations on “Data Fiduciaries”[23] processing personal data through automated or algorithmic systems. The Act ensures that algorithmic processing cannot obscure human liability.
- Information Technology Act, 2000 (and Intermediary Rules 2021)[24]: Sections 43A and 79 mandate due diligence and liability for digital intermediaries and automated system operators.
- Judicial and Policy Discourse: NITI Aayog’s National Strategy for Artificial Intelligence advocates responsible Al adoption (#AlforAll)[25] focused on ethical considerations, human oversight, and data sovereignty rather than conferring legal rights on algorithms.
LEGAL CHALLENGES AND DOCTRINAL ISSUES
Granting independent legal personhood to Al agents exposes foundational legal contradictions:
- Sanction and Enforcement Dilemmas: Traditional contractual and tortious principles, as discussed in judicial jurisprudence like United Steelworkers of America v American Manufacturing Co, [26]require clear consent and enforceable remedies. An algorithm cannot suffer incarceration or be deterred by conventional legal sanctions without financial capitalization requirement structures.
- Correlative Rights and Duties: Under Hohfeldian[27] legal theory, rights and duties are correlative. Because Al lacks consciousness, moral duties, or physical existence, conferring rights without enforceable duties disrupts basic legal equilibrium.
- Exploitation of Corporate Separation: Corporate actors could leverage autonomous software to evade tortious liabilities, making piercing the “Digital Corporate Veil”[28] extremely complex for judiciary bodies
CONCLUSION
The debate surrounding Al personhood reveals the boundary between technical autonomy and legal capacity. While algorithmic models perform complex tasks, granting them independent legal personhood remains doctrinally unviable and socially hazardous. As demonstrated by corporate, idol, and environmental precedents, legal fictions require human anchors to enforce duties and maintain accountability. Granting autonomous AI separate legal status without this anchor risks creating an insulation mechanism for developers and corporations. Global regulatory frameworks, including the EU Al Act 2024 and India’s DPDP Act 2023, correctly reject “electronic personhood” in favor of strict, human-centered liability frameworks. Legal systems must treat AI as a sophisticated tool under human oversight, ensuring technological innovation remains governed by accountability and the rule of law.
Author(s) Name: Shivani Gurjar (Prestige Institute of Management and Research Gwalior)
References:
[1] Margaret A Boden, Mind as Machine: A History of Cognitive Science (OUP 2006) vol 1, 45. (or vol 2, as applicable)
[2] Luciano Floridi, The Fourth Revolution: How the Infosphere is Reshaping Human Reality (OUP 2014) 112.
[3]Lawrence B Solum, ‘Legal Personhood for Artificial Intelligences’ (1992) 70 North Carolina Law Review 1231..
[4] Visa AJ Kurki, A Theory of Legal Personhood (OUP 2019) 25.
[5] Christopher D Stone, ‘Should Trees Have Standing?—Toward Legal Rights for Natural Objects’ (1972) 45 Southern California Law Review 450.
[6] P J Fitzgerald, Salmond on Jurisprudence (12th edn, Sweet & Maxwell 1966) 298.
[7] Friedrich Carl von Savigny, System des heutigen Römischen Rechts (Veit 1840) vol 2, 60.
[8] Otto von Gierke, Political Theories of the Middle Age (F W Maitland tr, Cambridge University Press 1900) 26.
[9] Rudolf von Jhering, Geist des römischen Rechts auf den verschiedenen Stufen seiner Entwicklung (Breitkopf und Härtel 1877) 341.
[10] John William Salmond, Jurisprudence or the Theory of the Law (Stevens and Haynes 1902) 350.
[11] Salomon v A Salomon & Co Ltd [1897] AC 22 (HL).
[12] Paul Davies and Sarah Worthington, Gower Principles of Modern Company Law (10th edn, Sweet & Maxwell 2016) ch 3.
[13] Gerhard Wagner, ‘Robot Liability: Click to Agree?’ (2019) 12 Journal of European Tort Law 27.
[14] Mohd Salim v State of Uttarakhand 2017 SCC OnLine Utt 367.
[15] Edwin Peel, Treitel on the Law of Contract (15th edn, Sweet & Maxwell 2020) para 2-001.
[16] Peter Watts and F M B Reynolds, Bowstead and Reynolds on Agency (22nd edn, Sweet & Maxwell 2021) art 1.
[17] Samir Chopra and Laurence F White, A Legal Theory for Autonomous Artificial Agents (University of Michigan Press 2011) 78.
[18] Stephen M Ulnick, ‘A Mechanical Tool Strategy for Algorithmic Automated Contracting’ (2021) 34 Harvard Journal of Law & Technology 411.
[19] David Vladeck, ‘Machines Without Principals: Liability Rules and Artificial Intelligence’ (2014) 89 Washington Law Review 117.
[20] Ryan Calo, ‘Robotics and the Lessons of Cyberlaw’ (2015) 124 Yale Law Journal 513
[21] John C Coffee Jr, ‘No Soul to Damn, No Body to Kick: An Unscandalized Inquiry into the Problem of Corporate Punishment’ (1981) 79 Michigan Law Review 386.
[22] European Parliament, Resolution of 16 February 2017 with Recommendations to the Commission on Civil Law Rules on Robotics (2015/2103(INL))
[23] Digital Personal Data Protection Act 2023, s 8.
[24] Information Technology Act 2000 (India), ss 43A, 79; Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Rules 2021.
[25] NITI Aayog, National Strategy for Artificial Intelligence: #AlforAll (Government of India 2018) 45.
[26] America v American Manufacturing Co, 363 US 564 (1960).
[27] Wesley Newcomb Hohfeld, ‘Some Fundamental Legal Conceptions as Applied in Judicial Reasoning’ (1913) 23 Yale Law Journal 16
[28] Jacob Turner, Robot Rules: Regulating Artificial Intelligence (Palgrave Macmillan 2019) 142.