Algorithmic Decision-Making And Corporate Criminal Liability In India - Re-Examining Fault Attribution In The AI Era

Author: LawSphereX

Category: Criminal Law

Published: 2026-08-23T17:47:26.959596+00:00

Updated: 2026-08-26T17:12:20.439+00:00

ALGORITHMIC DECISION-MAKING AND CORPORATE

CRIMINAL LIABILITY IN INDIA: RE-EXAMINING FAULT

ATTRIBUTION IN THE AI ERA

Nitish R Daniel, Assistant Professor of law at Barkatullah University and Ph.D Research scholar at National Law University Delhi.1

ABSTRACT

Rapid adoption of algorithmic decision-making is transforming how Indian corporations price products, screen borrowers, trade securities and manage supply chains. Yet when an autonomous system causes social harm through biased lending, errant trading or unsafe recommendations India’s centuryold fault attribution doctrines strain to respond. This article offers the first systematic re-examination of corporate criminal liability in India through the lens of artificial intelligence (AI). Part II maps four orthodox theories: vicarious liability, identification, aggregate knowledge and the corporatefault model and demonstrates why each rests on human agency assumptions that algorithmic autonomy disrupts.

Part III isolates three doctrinal pressure points unique to AI: the evidentiary collapse of mens rea, the tension between strict-liability offences and opacity-by-design, and the fragility of causal chains in self-learning systems. In Part IV, the analysis turns to India’s positive law through close reading of the Companies Act 2013, the Information Technology Act, and a cross-section of sector-specific regimes, which shows the inadequacy of present penal provisions to address AI harms. Statutes typically hinge liability on “persons in charge of, and responsible to, the company,” a formulation ill-suited where no natural person can reasonably foresee or control machine-generated outputs. Judicial glosses provide limited relief and risk inconsistent outcomes. Synthesizing these doctrinal and statutory gaps, the article advances a reform blueprint.

Key recommendations include: statutory duties of algorithmic governance (explainability, audit trails, bias testing) as predicates for compliance-based defences; a risk-tiered strict-liability matrix for high-impact deployments; calibrated reversal of the burden of proof on organisational due diligence; and safe-harbour incentives for open-source transparency. By recalibrating fault attribution around organisational risk management rather than elusive human intent, India can preserve criminal law’s deterrent force without stifling innovation in the AI era. 1 Assistant Professor of law at Barkatullah University and Ph. D Research scholar at National Law University Delhi.

I.

Iɴᴛʀᴏᴅᴜᴄᴛɪᴏɴ From banking to healthcare to transportation to human resources, artificial intelligence (AI) and algorithmic decision-making have drastically changed contemporary corporate operations. 2 From financial risk assessments and algorithmic trading to recruitment procedures and autonomous car operations, businesses are using AI systems to automate important choices more and more. 3 Corporate choices once taken by human managers and executives are increasingly being carried out either semi-autonomously or independently by AI-driven algorithms. Multinational companies such as Amazon, for example, greatly increase their operational efficiency and profitability by using AI-driven solutions for inventory management, shipping optimization, and tailored marketing.

4 Likewise, banking behemoths like JPMorgan Chase rely mostly on artificial intelligence algorithms for credit evaluations, algorithmic trading, and fraud detection, hence transforming conventional paradigms controlled by human judgment. 5

2 See M. Abbas Khan et al., Impact of Artificial Intelligence on the Global Economy and Technology

Advancements, 147, 147–180 In: ꜱ. ᴇʟ ʜᴀᴊᴊᴀᴍɪ ᴇᴛ ᴀʟ. , (ᴇᴅꜱ) ᴀʀᴛɪꜰɪᴄɪᴀʟ ɢᴇɴᴇʀᴀʟ ɪɴᴛᴇʟʟɪɢᴇɴᴄᴇ (ᴀɢɪ) ꜱᴇᴄᴜʀɪᴛʏ. ᴀᴅᴠᴀɴᴄᴇᴅ ᴛᴇᴄʜɴᴏʟᴏɢɪᴇꜱ ᴀɴᴅ ꜱᴏᴄɪᴇᴛᴀʟ ᴄʜᴀɴɢᴇʜᴇʟʟᴏ (Springer, Singapore)(2025); I. T Ayorinde & P. N. Idyorough, Exploring the frontiers of artificial intelligence: a comprehensive analysis, ɪɴɴᴏᴠᴀᴛɪᴏɴ ɪɴ ꜱᴄɪᴇɴᴄᴇ ᴀɴᴅ ᴛᴇᴄʜɴᴏʟᴏɢʏ, 35, 35-49 3. 4 (2024): J. Bughin, Artificial Intelligence, Its Corporate Use and How It Will Affect the Future of Work, 239, 239-260 In: L. Pᴀɢᴀɴᴇᴛᴛᴏ, (ᴇᴅs), Cᴀᴘɪᴛᴀʟɪsᴍ, Gʟᴏʙᴀʟ Cʜᴀɴɢᴇ ᴀɴᴅ Sᴜsᴛᴀɪɴᴀʙʟᴇ Dᴇᴠᴇʟᴏᴘᴍᴇɴᴛ (Springer, Cham)(2020). 3 See Noura Metawa et al. , Artificial Intelligence and Big Data for Financial Risk Management (Routledge) (2022); Bogojevic Arsic Vesna, Challenges of financial risk management: AI applications, Mᴀɴᴀɢᴇᴍᴇɴᴛ: Jᴏᴜʀɴᴀʟ ᴏғ Sᴜsᴛᴀɪɴᴀʙʟᴇ Bᴜsɪɴᴇss ᴀɴᴅ Mᴀɴᴀɢᴇᴍᴇɴᴛ Sᴏʟᴜᴛɪᴏɴs ɪɴ Eᴍᴇʀɢɪɴɢ Eᴄᴏɴᴏᴍɪᴇs 26. 3, 27, 27-34 (2021);.

Feichin Ted Tschang & Esteve Almirall, Artificial intelligence as augmenting automation: Implications for employment, Academy of Management Perspectives 35. 4, 642, 642-659(2021); Gil Cohen, Algorithmic trading and financial forecasting using advanced artificial intelligence methodologies, Mᴀᴛʜᴇᴍᴀᴛɪᴄs 10. 18, 3302 (2022); Iris H-Y. Chiu & Ernest WK Lim, Managing Corporations' Risk in Adopting Artificial Intelligence: A Corporate Responsibility Paradigm, Wᴀsʜ. U. Gʟᴏʙᴀʟ Sᴛᴜᴅ. L. Rᴇᴠ. 20, 347 (2021). 4 See Alex Kantrowitz, How Amazon Uses AI To Automate Work In Its Corporate Headquarters, Mᴇᴅɪᴜᴍ (Dec. 24, 2022)(https://kantrowitz. medium. com/how-amazon-uses-ai-to-automate-work-in-its-corporate-headquartersbbbc48e89769)(Last accessed Mar. 23, 2025); How Amazon Uses Artificial Intelligence? , SEASIA A CMMI ʟᴇᴠᴇʟ 5 ᴄᴏᴍᴘᴀɴʏ (Jan 03, 2023) (https://www. seasiainfotech.

com/blog/how-amazon-uses-artificial-intelligence/)(Last accessed Mar. 23, 2025); Amazon Staff, 8 ways Amazon is using generative AI to make life easier, from a more conversational Alexa to a better reviews experience, Aᴍᴀᴢᴏɴ (Dec. 19, 2023) (https://www. aboutamazon. com/news/innovation-at-amazon/how-amazon-uses-generative-ai) (Last accessed Mar. 23, 2025). 5 See K Tulsi et al. , Transforming Financial Services: The Impact of AI on JP Morgan Chase's Operational Efficiency and Decision-Making, International Journal of Scientific Research & Engineering Trends Volume 10 Issue 1, 207 (2024); J. P Morgan, How AI will make payments more efficient and reduce fraud, J. P. Mᴏʀɢᴀɴ (Nov. 20, 2023) (https://www. jpmorgan. com/insights/payments/payments-optimization/ai-payments-efficiency-fraudreduction)(Last accessed Mar.

23, 2025); Jeyadev Needhi, How AI Transformed Financial Fraud Detection: A Case Study of JP Morgan Chase (Jul. 8, 2024) (https://medium. com/@jeyadev_needhi/how-ai-transformed-financial-fraud-detection-a-case-study-of-jpmorgan-chase-f92bbb0707bb) (Last accessed Mar. 23, 2025).

Unquestionably efficient, cost-effective, and data-processing capable, AI-driven business decision-making also presents significant difficulties for assigning culpability, particularly in cases when algorithmic activities result in negative or criminal consequences. Originally meant to simplify recruiting by rapidly selecting the best candidates based on past hiring statistics, Amazon's AI system unintentionally acquired and spread current prejudices against women because of the previous majority of male employees in technological positions. 6 Resumes with signs of feminine identification so routinely scored worse, highlighting a discriminating result clearly connected to automated decision-making.

7 Uber's 2018 self-driving car disaster in Arizona, United States, claimed a pedestrian's life and spurred global discussions on corporate criminal responsibility in cases involving AI-driven mishaps. 8 Uber's criminal liability for negligent homicide was solely assigned to Rafaela Vasquez, the human vehicle operator, so absolving Uber from all criminal responsibility notwithstanding clear evidence showing Uber's faulty algorithmic technology, negligent operational policies, inadequate oversight mechanisms, and a poor safety culture. 9 Furthermore, Boeing's AI-driven automated control systems linked to disastrous Boeing 737 MAX disasters draw attention to serious issues regarding corporate responsibility when AI directly causes deaths. 10 These events underscore the hasty use of AI systems by organizations, emphasizing management decisions as primary catalysts for disastrous results.

Existing research has primarily focused on traditional human agency in corporate criminality, with limited attention given to the implications of AI-driven decisions within corporations. 11 6 See Gustavo Silveira Borges & Mauricio Da Cunha Savino Filo, Artificial Intelligence, Gender, and Human Rights: The Case of Amazon, 35 Rᴇᴠ. Jᴜsᴛ. Dɪʀᴇɪᴛᴏ 218 (2021).

7 See id.

8 See Troy Griggs and Daisuke Wakabaya,How a Self-Driving Uber Killed a Pedestrian in Arizona, Tʜᴇ Nᴇᴡ Yᴏʀᴋ Tɪᴍᴇs (Mar. 21, 2018) (https://www. nytimes. com/interactive/2018/03/20/us/self-driving-uber-pedestrian-killed. html) (Last accessed Mar. 27, 2025) 9 See Helen Stamp, The Reckless Tolerance of Unsafe Autonomous Vehicle Testing: Uber's Culpability for the Criminal Offense of Negligent Homicide, Case W. Res. JL Tech. & Internet 15, 37 (2024). 10 See William Langewiesche, What Really Brought Down the Boeing 737 Max? , Tʜᴇ Nᴇᴡ Yᴏʀᴋ Tɪᴍᴇs (Jul. 2, 2021)(https://www. nytimes. com/2019/09/18/magazine/boeing-737-max-crashes. html)((Last accessed Mar. 27, 2025) 11 See Soubhagya Ranjan Panda & Anita Sable, The Criminal Liability of Corporate Industries in India with Reference to USA & UK, Part 2 Indian J. Integrated Rsch. L.

2, 1 (2022); Titiksha Chhabra, Corporate Crime and the Criminal Liability of Corporate Entities with Regard to India: A Comparative Study, Part 2 Indian J. Integrated Rsch. L. 2, 1 (2022); Pratik Bharat Kashid, Corporate Criminal Liability, Part 2 Indian J. Integrated Rsch. L. 2, 1 (2022); Ramasayi Gummadi, Corporate Criminal Liability in India-A Theoretical Analysis, Nyaayshastra L. Rev. 2, 1 (2021); Rukmani Sachdeva, An Analytical Study on Corporate Criminal Liability: Comparative Study of India and USA, Issue 1 Int'l JL Mgmt. & Human. 5, 2577(2022); Mihailis E. Diamantis, The Extended Corporate Mind: When Corporations Use AI to Break the Law, 98 N. C. L. REV. 893 (2020); Mihailis E. Diamantis, Employed algorithms: a labor model of corporate liability for AI, Duke lJ 72, 797(2022).

This research aims to explore the complexities of corporate criminal liability in the context of AI integration within Indian corporations. Using a doctrinal legal approach, this paper examines legal doctrines, legislation and court rulings, to ascertain:

1. How do traditional theories of corporate criminal liability apply to offenses arising from

AI-driven decisions in Indian corporations?

2. To what extent does the current Indian legal framework address the accountability of

corporations for crimes involving AI technologies?

3. What legal reforms or policy measures could effectively bridge any gaps identified in

the existing framework to ensure robust corporate accountability for AI-induced offenses? It is contended that premature AI deployment, defined as the introduction of AI into operational settings prior to sufficient testing or risk mitigation, is not only a technological oversight but a significant managerial decision with far-reaching societal, ethical, and legal implications. Further, in India, corporate criminal responsibility is mostly regulated by judicial concepts, notably the “Doctrine of Identification,” which assumes the intent and direct involvement of a human actor. 12 Indian criminal jurisprudence has always prioritized mens rea (guilty thinking) and actus reus (guilty deed) of identifiable individuals. This conventional position becomes intricate when harm results from algorithmic systems lacking express human intent or direct human oversight.

Consequently, they seem inadequately prepared to manage liabilities arising from business acts autonomously performed by algorithms. The research article is structured into five key parts to systematically explore the complexities of corporate criminal liability in the context of AI-driven decision-making. Part II delves into the theoretical underpinnings of corporate criminal liability, reviewing traditional doctrines such as vicarious liability, the identification doctrine, and corporate fault theory. Part III critically examines how AI integration into corporate operations challenges the application of these traditional liability frameworks, highlighting issues such as the opacity of algorithmic decisions, the lack of human intent, and difficulties in establishing mens rea.

Part IV analyzes existing Indian legal provisions, including statutory clauses in sector-specific laws and relevant

12 See Iridium India Telecom Ltd. v. Motorola Inc., (2011) 1 SCC 74 (India).

judicial interpretations. Part V suggests necessary changes, offering concrete recommendations for adapting corporate criminal liability frameworks to the realities of AI-driven corporate conduct in India. II. Theories of Corporate Criminal Liability

Corporate immunity was the dominant view in early criminal law. This distrust originated from intellectual and legal issues. 13 First, corporations, being legal entities without a body, cannot be imprisoned, which are customary criminal sanctions. 14 Second, early legal philosophy held that criminal responsibility needed mens rea, which a corporation, as a non-sentient entity, cannot possess. 15 The argument was that companies cannot be held accountable for the criminal activities of their agents or employees due to the lack of vicarious liability in criminal law. 16 Third, the notion of ultra vires, which limited corporations to actions specifically approved by their constitutional texts, supported the idea that businesses could not commit crimes because crimes were never authorized corporate acts. 17 Thus, only individual corporate members with criminal intent could be charged.

The idea that a corporation might act willfully was deemed a legal fiction. 18 This viewpoint persisted throughout the 19th century and hindered corporate criminal jurisprudence. However, as corporate entities gained power and influence and began to have widespread social effects, particularly in public health, environmental damage, and financial fraud, courts and legislators eventually overcame these constraints. 19 This shift marked the birth of modern theories of corporate criminal liability, which sought to adapt the principles of criminal responsibility to the realities of organizational conduct and complexity. 13 See Kathleen F. Brickey, Corporate Criminal Accountability: A Brief History and an Observation, 60 Wᴀsʜ.

U. L.Q. 393, 396 (1981)

14 See Celia K. Wells, Corporate criminal liability: A ten year review, Cʀɪᴍ. LR12, 849, 849-878 (2014). 15 See John C. Coffee, “No soul to damn: no body to kick": An unscandalized inquiry into the problem of corporate punishment, Mɪᴄʜɪɢᴀɴ Lᴀᴡ Rᴇᴠɪᴇᴡ 79. 3, 386, 386-459 (1981); John C. Coffee Jr, Crime and the corporation: making the punishment fit the corporation, J. Cᴏʀᴘ. L. 47, 963(2021). 16 See James Gobert, Corporate criminality: four models of fault, Lᴇɢᴀʟ Sᴛᴜᴅɪᴇs 14. 3, 393, 393-410 (1994). 17 See L. H. Lᴇɪɢʜ, Tʜᴇ ᴄʀɪᴍɪɴᴀʟ ʟɪᴀʙɪʟɪᴛʏ ᴏғ ᴄᴏʀᴘᴏʀᴀᴛɪᴏɴs ɪɴ ᴇɴɢʟɪsʜ ʟᴀᴡ, 8-9 (Littlehampton Book Services Ltd) (1969); Trustees of Dartmouth College v. Woodward, 17 U. S. 518 (1819). 18 See Mihailis E. Diamantis, Corporate criminal minds, Nᴏᴛʀᴇ Dᴀᴍᴇ L. Rᴇᴠ. 91, 2049(2015); State v. Great Works Milling & Manufacturing Co. , 20 Me. 41 (1841). 19 See Wɪʟʟɪᴀᴍ S.

Lᴀᴜғᴇʀ, Cᴏʀᴘᴏʀᴀᴛᴇ ʙᴏᴅɪᴇs ᴀɴᴅ ɢᴜɪʟᴛʏ ᴍɪɴᴅs: Tʜᴇ ғᴀɪʟᴜʀᴇ ᴏғ ᴄᴏʀᴘᴏʀᴀᴛᴇ ᴄʀɪᴍɪɴᴀʟ ʟɪᴀʙɪʟɪᴛʏ, 9 (University of Chicago Press)(2006).

a. Vicarious Liability (Respondeat Superior) theory The first cases wherein corporations were held criminally liable involved public nuisance and offences. 20 Later it was extended to crimes that didn’t require mens rea. 21 Over time, the distinct legal personality of corporations, established in the civil sphere by cases such as Salomon v. Salomon, compelled jurists to ingeniously reconcile the disparity between human actions and corporate culpability. 22 By the early 20th century, courts started to apply principles from tort and agency law to criminal law, permitting the indictment of corporations. The U. S. Supreme Court's ruling in New York Central & Hudson River R. R. Co. v.

United States explicitly broadened the tort doctrine of respondeat superior to encompass criminal cases, rendering corporations accountable for offenses perpetrated by employees during the course of their employment and partially for the company's benefit. 23 Under the theory of vicarious liability (Respondeat Superior), a corporation is liable for the actions of its agents or employees, akin to a master bearing responsibility for the activities of a servant. 24 Its appeal is pragmatic, a company who gains from an employee's misconduct or fails to stop it can fairly be fined, introducing a due diligence defence i. e. , the corporation shows it took reasonable steps to prevent the offence, could help to offset this unfairness. The United States embraced this concept to a large degree, allowing businesses to be accountable for most crimes carried out under their purview.

25 Still, vicarious liability became popular particularly for regulatory violations i. e. , so-called strict liability or “no-fault” crimes. 26 This theory necessitates that firstly, a corporate agent (irrespective of their rank within the organisation) must first engage in an unlawful conduct (the actus reus) accompanied by the necessary mental condition (the mens rea). 27 Mens rea may be established through the ‘collective awareness’ of the employees as a collective entity, despite no one employee having

20 See R. v. Birmingham & Gloucester Ry. Co., (1842) 3 Q.B. 223 (Eng.).

21 See R. v. Great N. of Eng. Ry. Co., (1846) 9 Q.B. 315 (Eng.).

22 See Salomon v. A. Salomon & Co. Ltd., [1897] A.C. 22 (H.L.) (U.K.).

23 See New York Cent. & H.R.R. Co. v. United States, 212 U.S. 481 (1909).

24 See R.S.Welsh, The Criminal Liability of Corporations, LQ Rev.62, 345(1946).

25 See James R. Elkins, Corporations and the Criminal Law: An Un-easy Alliance, 65 Ky. LJ. 73, 87-88 (1976). 26 See Claire McIvor, The use and abuse of the doctrine of vicarious liability, Cᴏᴍᴍᴏɴ Lᴀᴡ Wᴏʀʟᴅ Rᴇᴠɪᴇᴡ 35.4,268, 268-296 (2006); James Jr, Fleming, Vicarious Liability, Tᴜʟ. L. Rᴇᴠ. 28,161 (1953). 27 See S. Sownthara Barani, Comparative Analysis of Corporate Criminal Liability, Issue 1 Indian JL & Legal Rsch. 5, 1(2023); James Gobert, The evolving legal test of corporate criminal liability, 61, 61-80 Lᴇᴏɴᴀʀᴅ Mɪɴᴋᴇs & Jᴏʜɴ Mɪɴᴋᴇs, Cᴏʀᴘᴏʀᴀᴛᴇ ᴀɴᴅ Wʜɪᴛᴇ Cᴏʟʟᴀʀ Cʀɪᴍᴇ (Sage Publications) (2008).

adequate information to recognise that a crime was occurring. 28 Secondly, the agent must have behaved within the scope of employment encompasses any action that transpired while the offending employee was engaged in a job-related activity. 29 Thirdly, the agent must have sought to advantage the corporation wherein the employee is not required to behave only for the advantage of the organisation, nor is the corporation obligated to obtain any actual profit. 30 Historically, numerous crimes emerged during the early industrial period, when the law was hesitant to place constraints on burgeoning enterprises. 31 As organisations have evolved in complexity and stability, this hesitance has diminished, leading to the expansion of vicarious liability frameworks. 32 Nonetheless, the conflict between practicality and principle endures.

It is virtually unfeasible for senior executives to monitor every move of each employee. Conversely, public policy necessitates that firms implement stringent supervision and proactive governance in the face of anticipated risks of harm. 33 Judicial decisions show that even where the actus reus is satisfied through vicarious attribution, courts often hesitate to impute mens rea, limiting vicarious liability to strict liability offences or cases where the employee is specifically entrusted with statutory duties. 34 Although successful in implicating companies in legal actions, vicarious liability frequently faces criticism for its excessive scope and moral randomness. A corporation with an exemplary compliance record may still bear criminal liability for the actions of a rogue employee acting independently against corporate policy.

35 Critics contend that this represents 28 See Anthony Ragozino, Replacing the collective knowledge doctrine with a better theory for establishing corporate mens rea: The duty stratification approach, Sᴡ. UL Rᴇᴠ. 24, 423(1994). 29 See Alan O. Sykes, The boundaries of vicarious liability: An economic analysis of the scope of employment rule and related legal doctrines, Harvard Law Review 101. 3, 563, 563-609(1988). 30 See Harold J. Laski, Basis of vicarious liability, Yᴀʟᴇ LJ 26, 105 (1916); Doug Cassel, Corporate aiding and abetting of human rights violations: Confusion in the courts, Nᴡ. UJ Iɴᴛ'ʟ Hᴜᴍ. Rᴛs. 6, 304 (2007). 31 See David Philips, Crime, law and punishment in the Industrial Revolution, In: Pᴀᴛʀɪᴄᴋ O'Bʀɪᴇɴ & Rᴏʟᴀɴᴅ Qᴜɪɴᴀᴜʟᴛ (Eᴅs. ), Tʜᴇ Iɴᴅᴜsᴛʀɪᴀʟ Rᴇᴠᴏʟᴜᴛɪᴏɴ ᴀɴᴅ Bʀɪᴛɪsʜ Sᴏᴄɪᴇᴛʏ, 156, 156-83 (Cambridge University Press)(1993).

32 See id.

33 See Steven P. Croley, Vicarious liability in tort: On the sources and limits of employee reasonableness, S. Cᴀʟ. L. Rᴇᴠ. 69, 1705(1995); William S.Laufer, & Alan Strudler, Corporate intentionality, desert, and variants of vicarious liability, Aᴍ. Cʀɪᴍ. L. Rᴇᴠ. 37 1285 (2000). 34 See Seaboard Offshore Ltd. v. Sec’y of State for Transp., [1993] 1 W.L.R. 1025 (U.K.); Nat’l Rivers Auth. (S. Region) v. Alfred McAlpine Homes E. Ltd., [1994] EWHC J0126-9 (U.K.); Alphacell Ltd. v. Woodward, [1972] A.C. 824 (H.L.) (U.K.).

35 See Evan Tuttle, Reexamining the Vicarious Criminal Liability of Corporations for the Willful Crimes of Their Employees, Cʟᴇᴠ. Sᴛ. L. Rᴇᴠ. 70, 121 (2021): Robert Luskin, Caring about Corporate 'Due Care': Why Criminal Respondeat Superior Liability Outreaches Its Justification, Aᴍ. Cʀɪᴍ. L. Rᴇᴠ. 57, 303(2020); Preet Bharara, Corporations cry uncle and their employees cry foul: rethinking prosecutorial pressure on corporate defendants, Aᴍ. Cʀɪᴍ. L. Rᴇᴠ. 44, 53 (2007); Albert W. Alschuler, Two ways to think about the punishment of corporations, Aᴍ. Cʀɪᴍ. L. Rᴇᴠ. 46, 1359 (2009).

strict liability disguised as criminal law, imposing penalties without genuine culpability. The counter-argument contends that corporations face no possibility of imprisonment, only financial penalties, rendering vicarious liability an essential economic deterrent that incentivises companies to enhance employee supervision and training. 36 To strike a balance, many commentators advocate for a due diligence defence, allowing corporations to avoid liability by demonstrating that they took all reasonable steps to prevent the offence. 37 b. The Identification (Alter Ego) Theory The identification (Alter ego) theory posits that a corporation is directly accountable solely for the actions and intentions of senior officials who embody the company's “mind”.

38 This concept begins by identifying a real person, usually a director or senior executive, who perpetrated the crime, and subsequently assesses whether that individual possesses sufficient stature to be considered the corporation's alter ego. 39 In such cases, the individual's mens rea and actions are considered as those of the organisation. 40 In Lennard’s Carrying Co. it was observed that “A corporation is an abstract. It has no mind of its own any more than it has a body of its own, its acting and directing will must consequently be sought in the person of somebody ... who is really the directing mind or will of the corporation, the very ego and centre of personality of the corporation. ”41 In H. L. Bolton (Engg. ) Co. Ltd. v. T. J. Graham & Sons Ltd.

,42 Denning LJ compared a corporation to a human with a brain and limbs wherein workers, servants, and agents are simply tools for a corporation (the limbs), whereas directors and managers are the decision-makers and steer the company (brain). The decision implies that a corporation can be held guilty of 36 See Albert W. Alschuler, Two ways to think about the punishment of corporations, Aᴍ. Cʀɪᴍ. L. Rᴇᴠ. 46, 1359 (2009). 37 See Roni A. Elias, The virtues of the due diligence defense for corporations in criminal cases: solving the problems of a corporation's vicarious liability for the crimes of its agents and employees, Gᴇᴏ. JL & Pᴜʙ. Pᴏʟ'ʏ 13, 423 (2015). 38 See Antoinette Sedillo Lopez, The Alter Ego Doctrine: Alternative Challenges to the Corporate Form, UCLA L. Rᴇᴠ. 30, 129(1982); Elizabeth A. Evans & Daniel G. Lentz, Alter Ego, In: Rᴏᴍᴀɴ L. Wᴇɪʟ ᴇᴛ ᴀʟ.

(Eᴅs), Lɪᴛɪɢᴀᴛɪᴏɴ Sᴇʀᴠɪᴄᴇs Hᴀɴᴅʙᴏᴏᴋ: Tʜᴇ Rᴏʟᴇ ᴏғ ᴛʜᴇ Fɪɴᴀɴᴄɪᴀʟ Exᴘᴇʀᴛ, 1, 1-21(2017).

39 See id.

40 See R. v. I.C.R. Haulage Ltd., [1944] K.B. 551 (C.C.A.) (U.K.); Tesco Supermarkets Ltd. v. Nattrass, [1972] A.C. 153 (H.L.) (U.K.).

41 Lennard's Carrying Co. Ltd. v. Asiatic Petroleum Co. Ltd. 1915 AC 705, 715 (HL).

42 See HL Bolton (Eng’g) Co. v. T.J. Graham & Sons Ltd., [1957] 1 Q.B. 159 (U.K.).

offences requiring a guilty mind if the person or group in command of the company's affairs at the time of the offence can be identified. This notion became known as the “identification principle”. 43 It was observed that: “A company may in many ways be likened to a human body. They have a brain and a nerve centre which controls what they do. They also have hands which hold the tools and act in accordance with directions from the centre. Some of the people in the company are mere servants and agents who are nothing more than hands to do the work and cannot be said to represent the mind or will. Others are directors and managers who represent the directing mind and will of the company, and control what they do. The state of mind of these managers is the state of mind of the company and is treated by the law as such.

So you will find that in cases where the law requires personal fault as a condition of liability in tort, the fault of the manager will be the personal fault of the company. That is made clear in Lord Haldane's speech in Lennard's Carrying Co. Ltd. v. Asiatic Petroleum Co. l,. fd. (AC at pp. 713, 714). So also in the criminal law, in cases where the law requires a guilty mind as a condition of a criminal offence, the guilty mind of the directors or the managers will render the company themselves guilty. ”44 Explaining the same MacNaghten, J. in Director of Public Prosecutions v. Kent and Sussex Contractors Ltd. 45 observed: “A body corporate is a "person" to whom, amongst the various attributes it may have, there should be imputed the attribute of a mind capable of knowing and forming an intention - indeed it is much too late in the day to suggest the contrary.

It can only know or form an intention through its human agents, but circumstances may be such that the knowledge of the agent must be imputed to the body corporate. Counsel for the respondents says that, although a body corporate may be capable of having an intention, it is not capable of having a criminal intention. In this particular case the intention was the intention to deceive. If, as in this case, the responsible agent of a body 43 See John Mugambwa et al. ,. Corporate Liability, In: Cᴏᴍᴍᴇʀᴄɪᴀʟ ᴀɴᴅ Bᴜsɪɴᴇss Oʀɢᴀɴɪᴢᴀᴛɪᴏɴs Lᴀᴡ ɪɴ Pᴀᴘᴜᴀ Nᴇᴡ Gᴜɪɴᴇᴀ, 429, 429-523 (Routledge-Cavendish)(2007); Davide Ravasi & Johan Van Rekom, Key issues in organizational identity and identification theory, Cᴏʀᴘᴏʀᴀᴛᴇ Rᴇᴘᴜᴛᴀᴛɪᴏɴ Rᴇᴠɪᴇᴡ 6, 118, 118-132(2003); Tesco Stores Ltd v. Brent London Borough Council Citation: [1993] 1 WLR 1037 (House of Lords)

44 See id at AC p. 172.

45 See Dir. of Pub. Prosecutions v. Kent & Sussex Contractors Ltd., [1944] K.B. 146 (U.K.).

corporate puts forward a document knowing it to be false and intending that it should deceive, I apprehend, according to the authorities that Viscount Caldecote, L. C. J. , has cited, his knowledge and' intention must be imputed to the body corporate. ”46 Highlighting the difference from vicarious liability the House of Lords in Tesco E Supermarkets Ltd. v. Nattrass47 observed: “... A living person has a mind which can have knowledge or intention or be negligent and he has hands to carry out his intentions. A corporation has none of these: it must act through living persons, though not always one or the same person. Then the person who acts is not speaking or acting for the company. He is acting as the company and his mind which directs his acts is the mind of the company. There is no question of the company being vicariously liable.

He is not acting as a servant, representative, agent or delegate. He is an embodiment of the company or, one could say, he hears and speaks through the persona of the company, within his appropriate sphere, and his mind is the mind of the company. If it is a guilty mind then that guilt is the guilt of the company. It must be a question of law whether, once the facts have been ascertained, a person in doing particular things is to be regarded as the company or merely as the company's servant or agent. In that case any liability of the com~any can only be a statutory or vicarious liability. ”48 The House of Lords also determined that a company can only be held accountable for the actions of its senior executives, directors or officers, whose judgements effectively embody the firm.

In Tesco, the mispricing blunder of a branch manager was not attributed to the company, as he was considered too low in the hierarchy to be regarded as its governing mind. 49 Identification thereby establishes a “limited subset” of individuals whose actions may implicate the organisation. This method prevents holding organisations accountable for minor incidents, however it has faced criticism for allowing major corporations to evade culpability through decentralised decision-making. 50 In intricate organisations, culpability for corporate

46 See id at AC p. 156.

47 See Tesco Supermarkets Ltd. v. Nattrass, [1972] A.C. 153 (H.L.) (U.K.).

48 See id at AC p. 170 E-G.

49 See id.

50 See All Answers ltd, Corporate Manslaughter and Corporate Homicide Act, Lᴀᴡᴛᴇᴀᴄʜᴇʀ.ɴᴇᴛ (April

2025)(https://www.lawteacher.net/free-law-essays/criminal-law/corporate-manslaughter-and-corporatehomicide-act.php?vref=1)(Last accessed Apr. 13, 2025)

crime is never attributable to one individual; rather, it may be disseminated across a vicepresident, committees, middle managers, supervisors, and online employees, rendering it difficult to assign blame to one specific person. 51 Conversely, it is over-inclusive where directors or officers commit crimes contrary to company policy, yet their status alone is enough to implicate the company. In Moore v. Bresler, the courts affirmed corporate liability for fraudulent tax returns filed by top executives, despite the activities being intended to obscure personal wrongdoing rather than to advantage the corporation. 52 The court indicated that formal roles might supersede intent in assessing corporate accountability under this framework.

In R v P&O Ferries53, a roll-on/roll-off ferry owned by Townsend Thoresen crashed shortly after departure from Zeebrugge on March 6, 1987, killing 193 people. The ferry left the harbour with its bow doors open, flooding the vehicle deck and causing it to capsize. The assistant boatswain who locked the doors slept in his cabin. Insufficient communication, lack of door closure rules, and pervasive safety culture issues plagued the business. The prosecution found no senior director or equivalent with mens rea in the Herald of Free Enterprise case. Despite extensive corporate irresponsibility, no top official could be legally named the company's “mind” for gross negligence homicide.

Similarly, the Ladbroke Grove train incident also known as the Paddington rail crash which occurred on 5 October 1999 demonstrated the inadequacy of the identification concept in assigning criminal liability to companies for homicide; wherein a Thames train operator ignored a red signal and hit a high-speed train head-on, killing 31 and injuring over 400. 54 Investigations uncovered numerous systematic deficiencies in training, signal visibility, risk management, and corporate safety protocols across various entities, including Thames Trains, Railtrack (the infrastructure operator), and contractors; yet no directing mind with mens rea existed. 51 See Ephraim N. Ngwafor, Re-Visitation of the Alter Ego Doctrine in Corporate Criminal Liability, Kɪɴɢsᴛᴏɴ L. Rᴇᴠ. 13, 3 (1983); , Kineisha Salma Syakira et al.

, The Existence of Individual Companies As Legal Entities In Micro and Small Enterprises In Indonesia: A Review of The Problems of The Alter Ego Doctrine, Jᴜʀɴᴀʟ Iʟᴍᴜ Hᴜᴋᴜᴍ Tᴀᴍʙᴜɴ Bᴜɴɢᴀɪ 9. 1, 450, 450-461 (2024).

52 See Moore v. Bresler, [1944] 2 All E.R. 515 (U.K.).

53 See R. v. P&O Ferries, [1991] 93 Cr. App. R. 72 (U.K.).

54 See Paul Clifton, Lessons from Ladbroke Grove crash 25 years on, BBC (Oct. 4, 2024)

(https://www.bbc.com/news/articles/cje3kd8y1j0o) (Last accessed Apr. 13, 2024)

Thus, the doctrine promotes fragmented management structures, delegating operational control while narrowly defining legal responsibility. Consequently, criminal liability frequently does not arise, even when organisational culture or policy indirectly promotes unlawful behaviour. To address this gap, the UK enacted the Corporate Manslaughter and Corporate Homicide Act 2007, under which for the first time, a corporation could be convicted of corporate manslaughter or corporate homicide based on “management failures” that amount to a gross breach of a duty of care, causing death. Another modern development in UK law is the creation of “failure to prevent” offences, which impose liability on companies for failing to stop certain crimes by their associates.

55 The UK Bribery Act 2010 introduced a strict liability offence for companies that fail to prevent bribery by persons associated with the business and a company can avoid conviction only by proving it had “adequate procedures” in place to prevent bribery. 56 Similarly, the Criminal Finances Act 2017 holds companies liable for failing to prevent facilitation of tax evasion. 57 These offences mark a shift towards an organizational duty to prevent wrongdoing; effectively penalizing a form of corporate omission or negligence in oversight. c. Aggregate theory To resolve the aforementioned issue, certain courts and pundits have suggested aggregation, a process by which the knowledge and conduct of multiple individuals are amalgamated to assign accountability to the business i. e. , the corporation's “mens rea” could be established by consolidating information possessed by several agents.

58 For example, one employee's awareness of a safety issue combined with another employee's inaction may collectively constitute the corporation's intentional infraction. 59 U. S. courts have intermittently utilised a variant of the “collective knowledge” theory, as exemplified in US v. Bank of New England, where a bank's knowledge was considered the aggregate of its workers' knowledge. 60 Many jurisdictions, however, have approached this theory with caution; English courts predominantly rejected aggregation as a foundation for criminal liability, particularly in R. v. HM Coroner for

55 Corporate Manslaughter and Corporate Homicide Act, 2007, No. 19, Act of Parliament, 2007 (UK).

56 Bribery Act, 2010, No. 23, Act of Parliament, 2010 (UK).

57 Criminal Finances Act, 2017, No. 22, Act of Parliament, 2017 (UK).

58 See Virginia Harper Ho, Theories of corporate groups: corporate identity reconceived, Sᴇᴛᴏɴ Hᴀʟʟ L. Rᴇᴠ. 42, 879(2012); David Millon, Theories of the Corporation, Dᴜᴋᴇ LJ, 201, 210-213 (1990).

59 See id.

60 See United States of America v. Bank of New England 821 F.2d 844 (1st Cir. 1987).

East Kent, ex parte Spooner. 61 The concern is that it can be too easy to hold a company liable under this doctrine. The aggregation model illustrates its effectiveness in incidents such as the Clapham rail crash. 62 Individual lapses, including a technician's wiring error and managerial oversight failures, collectively resulted in tragedy. Within an aggregate framework, corporate liability does not hinge on pinpointing a specific employee responsible for both actus reus and mens rea; rather, it focuses on whether British Rail, as an institution, possessed adequate knowledge and capability to avert the harm. However, aggregation is not without criticism. It raises fairness concerns, as individuals whose actions are collectively attributed to the corporation may lack the requisite culpability for personal liability.

The risk lies in constructing corporate liability from discrete, nonblameworthy acts, thereby potentially inflating criminal penalties. Aggregating knowledge or conduct can be useful in addressing complex offences, such as frauds involving multiple actors, but it invites the question: how much aggregation is too much? Should a firm be held liable if four separate employees each negligently overlooked a red flag, even though none individually acted with gross negligence? This may undermine due process, as a corporation could be convicted where all individuals would be acquitted if tried separately. Moreover, successful prosecution under a cultural fault model demands that enforcement agencies possess the resources and expertise to prove systemic deficiencies,a task far more difficult than demonstrating individual mens rea. d.

Corporate Fault Model As previously discussed, determining how a company can be “at fault” has long posed challenges for criminal jurisprudence. Traditional doctrines such as identification and vicarious liability attempt to bridge this conceptual gap by attributing individual acts and mental states to the corporation. Yet these approaches often fall short, particularly in complex or AI-mediated contexts, because they rely on individual culpability even when the misconduct arises from systemic organizational failures.

61 See R. v. HM Coroner for E. Kent ex parte Spooner, (1989) 88 Cr. App. R. 10 (U.K.).

62 See Gʀᴇɢ Mᴏʀsᴇ, Tʜᴇ Cʟᴀᴘʜᴀᴍ Tʀᴀɪɴ Aᴄᴄɪᴅᴇɴᴛ: Cᴀᴜsᴇs, Cᴏɴᴛᴇxᴛ ᴀɴᴅ ᴛʜᴇ Cᴏʀᴘᴏʀᴀᴛᴇ Mᴇᴍᴏʀʏ Cʜᴀʟʟᴇɴɢᴇ

(Pen and Sword Transport)(2023).

The concept of corporate fault offers a more pragmatic and philosophically grounded alternative. Corporate policy rarely originates with a single individual; it develops through committees, internal deliberation, managerial endorsement, and board approval. This process generates a synthesized institutional will that is distinct from any single employee and properly attributable to the firm itself. Moving beyond derivative liability, whether traced to one individual or dispersed among many, the organizational fault model contends that a company may be culpable by virtue of its deficient culture or policies. 63 In this context, responsibility is not contingent upon identifying any specific culpable individual; the emphasis is on whether the corporation's structures, practices, or culture facilitated or condoned the offence.

64 James Gobert articulates this as identifying the blameworthiness that underpins criminal liability within the corporation itself rather than attributing it to any one agent. 65 A corporation that consistently neglects safety regulations or implicitly endorses unethical financial practices may be perceived as possessing an organisational mens rea. 66 This comprehensive approach regards the corporation as an autonomous entity with distinct objectives and characteristics. 67 It corresponds with the intuitive concept that Ford or IBM can possess a distinct identity independent of any individual employee. Organisational theory underpins contemporary notions such as corporate culture as indicators of deficiencies and has impacted legislative reforms in certain nations. 68 It also encompasses offences of omission (neglecting to avert misconduct).

69 This paradigm does not supplant individual culpability; instead, both the corporation and responsible individuals may be prosecuted, illustrating that corporate crime frequently entails a confluence of corporate ethos and human acts. This theory is especially relevant in instances of dispersed omissions, where damage arises from numerous little failures instead of a singular, identifiable action. The Bhopal Gas Tragedy of December 1984 exemplifies corporate negligence, predating the formal adoption 63 See Neil Cavanagh, Corporate criminal liability: an assessment of the models of fault, The Journal of Criminal Law 75. 5, 414, 414-440(2011); Richard Mays, Towards corporate fault as the basis of criminal liability of corporations, Mountbatten Journal of Legal Studies 2.

2, 31, 31-67 (1998); Brent Fisse, Reconstructing corporate criminal law: Deterrence, retribution, fault, and sanctions, S. Cᴀʟ. L. Rᴇᴠ. 56, 1141(1982).

64 See id.

65 See James Gobert, Corporate criminality: four models of fault, Lᴇɢᴀʟ Sᴛᴜᴅɪᴇs 14.3, 393, 407-408 (1994). 66 See Vikramaditya S. Khanna, Is the notion of corporate fault a faulty notion: The case of corporate mens rea, BUL ʀᴇᴠ. 79, 355(1999).

67 See id.

68 See Tᴏᴍ Cʜʀɪsᴛᴇɴsᴇɴ ᴇᴛ ᴀʟ., Oʀɢᴀɴɪᴢᴀᴛɪᴏɴ ᴛʜᴇᴏʀʏ ᴀɴᴅ ᴛʜᴇ ᴘᴜʙʟɪᴄ sᴇᴄᴛᴏʀ: Iɴsᴛʀᴜᴍᴇɴᴛ, ᴄᴜʟᴛᴜʀᴇ ᴀɴᴅ ᴍʏᴛʜ (Routledge) (2020). 69 See Robert Sullivan, Conduct and Complicity: Liability Based on Omission and Risk, Cᴀᴍʙʀɪᴀɴ L. Rᴇᴠ. 39, 60 (2008); Otto Kirchheimer, Criminal omissions, Hᴀʀᴠ. L. Rᴇᴠ. 55, 615(1941); Arthur Leavens, A Causation Approach to Criminal Omissions, Cᴀʟɪғ. L. Rᴇᴠ. 76, 547(1988).

of the phrase in criminal law. 70 The emission of methyl isocyanate (MIC) gas from the Union Carbide India Ltd (UCIL) facility, a subsidiary of the U. S. -based Union Carbide Corporation (UCC), resulted in the fatalities of over 10,000 individuals within days and inflicted enduring health issues and environmental destruction that continues to this day. Investigations disclosed that UCC had been cognisant, as early as 1982, of the plant's significant safety deficiencies and the anticipated possibility of an uncontrolled response. 71 Notwithstanding this, safety measures were disregarded, suggested repairs were deferred, and essential systems remained nonfunctional; decisions motivated by financial constraints and administrative indifference.

72 These findings illustrate that the harm originated not from individual recklessness, but from a systematic failure in organisational governance and risk management. However, criminal liability was disproportionately borne by Indian officials of UCIL, while UCC’s American leadership, including CEO Warren Anderson, escaped punishment. 73 The global framework of UCC enabled it to capitalise on jurisdictional voids, illustrating the risks associated with "legal distance" in transnational corporate activities. This tragedy highlights the inadequacy of conventional doctrines such as identification or vicarious culpability, which rely on attributing guilt to particular individuals.

Rather, it robustly endorses the transition to a corporate fault model, wherein the legal emphasis is on whether a corporation, as a decision-making body, neglected to establish sufficient compliance procedures, disregarded known risks, and permitted practices likely to cause harm. In a modern setting when AI systems operate with unclear reasoning and dispersed oversight, Bhopal gas tragedy stands as a cautionary example highlighting that complexity and decentralisation must not exempt businesses from accountability when avoidable damages transpire under their supervision.

Having examined the four principal theories of corporate criminal liability, vicarious liability, identification theory, aggregation, and corporate fault, it becomes evident that each is grounded 70 See Usha Ramanathan, The Bhopal case: retrospect and prospect, In: 138, 138-145, Pʜɪʟɪᴘᴘᴇ Cᴜʟʟᴇᴛ & Sᴜᴊɪᴛʜ Kᴏᴏɴᴀɴ (Eᴅs), Rᴇsᴇᴀʀᴄʜ Hᴀɴᴅʙᴏᴏᴋ ᴏɴ Lᴀᴡ, Eɴᴠɪʀᴏɴᴍᴇɴᴛ ᴀɴᴅ ᴛʜᴇ Gʟᴏʙᴀʟ Sᴏᴜᴛʜ (Edward Elgar Publishing) (2019); Roli Varma & Daya R. Varma, The Bhopal disaster of 1984, Bulletin of Science, Technology & Society 25. 1, 37, 37-45(2005). 71 See Tim Covell, Bhopal Disaster Litigation: It's Not over Yet, Tʜᴇ NCJ Iɴᴛ'ʟ L. & Cᴏᴍ. Rᴇɢ. 16, 279 (1991); Trotter, R. Clayton et al. , Bhopal, India and union carbide: The second tragedy, Journal of Business Ethics 8, 439, 439-454 (1989); Sudhir K.

Chopra, Multinational Corporations in the Aftermath of Bhopal: The Need for a New Comprehensive Global Regime for Transnational Corporate Activity, Vᴀʟ. UL Rᴇᴠ. 29, 235(1994).

72 See id.

73 See Tim Edwards, Criminal failure and" the chilling effect": a short history of the Bhopal criminal prosecutions, Social Justice 41.1⁄2 (135-136), 53, 53-79 (2014); G. S. Bajpai & Bir Pal Singh, The Bhopal Gas Disaster and Corporate Criminal Negligence, In: Mᴀɴɢᴀɪ Nᴀᴛᴀʀᴀᴊᴀɴ, Iɴᴛᴇʀɴᴀᴛɪᴏɴᴀʟ Cʀɪᴍᴇ ᴀɴᴅ Jᴜsᴛɪᴄᴇ, 200 (Cambridge University Press) (2011).

in foundational assumptions about human agency, intent, and control within corporate structures. These doctrines emerged in an era when harmful corporate conduct could generally be traced to decisions made by identifiable individuals operating within a clear hierarchy. Today, however, the corporate environment is undergoing a profound transformation with the integration of AI systems into decision-making processes. Part II explores the complex challenges introduced by AI, particularly with respect to intent, causation, autonomy, and opacity, and critically evaluates how traditional theories of corporate criminality either struggle or adapt when applied to algorithmic actors III. AI and the Doctrinal Disruption

a. Problem of Mens rea A fundamental component of the majority of criminal acts is the necessity of a guilty mind

(mens rea). AI systems at the current stage of technological development, especially those employing machine learning, deep learning, or neural networks; lack consciousness, emotions, or intentions as understood in a human context. 74 The criminal law requires various forms of mens rea i. e. , actus reus was performed intentionally (“To Intend is to have in mind a fixed purpose to reach a desired objective... denote(s) the state of mind of a man who not only foresees but also desires the possible consequences of his conduct”),75 with knowledge (“...

signify a state of mental realization with the bare state of conscious awareness of certain facts in which human mind remains supine or inactive”,76 dishonestly (“means doing anything with the intention of causing wrongful gain to one person or wrongful loss to another person”),77 fraudulently (“means doing anything with the intention to defraud but not otherwise”,78 voluntarily (... “whereby he intended to cause it, or by means which, at the time 74 Sadia Tariq et al. , Examining Some Serious Challenges and Possibility of AI Emulating Human Emotions, Consciousness, Understanding and ‘Self’, Journal of NeuroPhilosophy 1. 1, 53, 53-72 (2022); Zhenhua Zhou, Emotional thinking as the foundation of consciousness in artificial intelligence, Cultures of Science 4. 3, 112, 112-123 (2021); 75 See Jai Prakash v. State (Delhi Admin.

), (1991) 2 SCC 32 (India); Bharatiya Nyaya Sanhita, 2023, § 324, No. 45, Acts of Parliament, 2023 (India).

76 See id; Imtiaz Yusufbhai Shaikh v. State of Gujarat, R/CR.MA/14919/2015 (Guj. HC) (India).

77 See Bharatiya Nyaya Sanhita, 2023, § 2 (7) No. 45, Acts of Parliament, 2023 (India); Bharatiya Nyaya Sanhita, 2023, § 316; § 318–319, No. 45, Acts of Parliament, 2023 (India); Queen Empress v. Agha Muhammad Yusuf 1895 SCC Online All 87; Queen Empress V. Churn Chunga 1895 SCC Online Cal 13; Budhan Singh v. State 1960 SCC Online Pat 66. 78 See Bharatiya Nyaya Sanhita, 2023, § 2 (9), 320–322, No. 45 Acts of Parliament, 2023 (India); S. Dutt v. State of U.P., AIR 1966 SC 523 (India); Queen Empress v. Soshi Bhushan, 1893 SCC OnLine All 25 (India); Vimla v. Delhi Admin., AIR 1963 SC 1572 (India); In re London & Globe Fin. Corp. Ltd., [1903] 1 Ch. 728 (U.K.).

of employing those means, he knew or had reason to believe to be likely to cause it. ”). 79 Therefore, the legal doctrines of intent and causation, foundational to attributing criminal liability in common law systems, are based on human behavior and mental states. 80 So, even if AI advances to the stage of independent intent or thoughts, the current law would not interpret it as capable of having mens rea. So, the thought of imputing mens rea to AI isn’t feasible. So the next logical question is can the human developing the AI be imputed to the mens rea of AI? However, the doctrine encounters substantial limitations when applied to fully autonomous or machine-learning-based AI systems that make decisions independent of direct human input.

This is especially problematic when AI-generated conduct arises from unanticipated algorithmic behavior; a scenario commonly referred to as the “black box” problem. 81 AI systems, particularly those utilising deep neural networks, function as opaque entities, even to their developers. 82 Their internal logic cannot be readily reverse-engineered, elucidated, or comprehended retrospectively. 83 The philosophical analysis of action and responsibility further exposes the doctrinal inadequacy of attributing traditional mens rea to AI systems. Intent necessitates volition or, at a minimum, the anticipation of harm. As Robert Audi emphasizes, true human agency requires not only the presence of intentions but also volitions; momentary acts of will that causally structure bodily movements and decisions.

84 Audi distinguishes between mere outcomes and actions that are mentally authored through volitional processes, arguing that volition is a necessary component for grounding moral and legal responsibility. 85 In contrast, AI systems, particularly those 79 See Bharatiya Nyaya Sanhita, 2023, §2(33), No. 45 Acts of Parliament, 2023 (India); Meeru Bhatia Prasad v. State, 2001 SCC OnLine Del 1101 (India); Queen Empress v. Raghunath Rai, 1892 SCC OnLine All 59 (India). 80 See Yavar Bathaee, The Artificial Intelligence Black Box and the Failure of Intent and Causation, 31 Hᴀʀᴠ. J. L. & Tᴇᴄʜ 889 (Spring 2018).

81 See Warren J. Von Eschenbach, Transparency and the black box problem: Why we do not trust AI, Philosophy & Technology 34. 4, 1607, 1607-1622 (2021); Carlos Zednik, Solving the black box problem: A normative framework for explainable artificial intelligence, Philosophy & technology 34. 2, 265, 265-288 (2021); Charlotte A. Tschider, Beyond the “Black Box”, Dᴇɴᴠ. L. Rᴇᴠ. 98, 683 (2020); Vladislav V. Fomin & Paulius Astromskis, The Black Box Problem In: Jᴏʜɴ-Sᴛᴇᴡᴀʀᴛ Gᴏʀᴅᴏɴ (Eᴅ), Fᴜᴛᴜʀᴇ Lᴀᴡ, Eᴛʜɪᴄs, ᴀɴᴅ Sᴍᴀʀᴛ Tᴇᴄʜɴᴏʟᴏɢɪᴇs, 112, 112-125 (Brill) (2023). 82 See Amina Adadi & Mohammed Berrada, Peeking inside the black-box: a survey on explainable artificial intelligence (XAI), IEEE access 6, 52138, 52138-52160(2018); Jean-Christophe Bélisle-Pipon et al.

, Artificial intelligence ethics has a black box problem, AI & Sᴏᴄɪᴇᴛʏ, 1, 1-16 (2023); Yavar Bathaee, The artificial intelligence black box and the failure of intent and causation, Hᴀʀᴠ. JL & Tᴇᴄʜ. 31, 889 (2017); Warren J. Von Eschenbach, Transparency and the black box problem: Why we do not trust AI, Philosophy & Technology 34. 4, 1607, 1607-1622 (2021). 83 See Carlos Zednik, Solving the black box problem: A normative framework for explainable artificial intelligence, Pʜɪʟᴏsᴏᴘʜʏ & ᴛᴇᴄʜɴᴏʟᴏɢʏ 34. 2, 265 (2021)

84 See Robert Audi, Volition, Intention, and Responsibility, U. Pᴀ. L. Rᴇᴠ. 142, 1675(1993).

85 See id at 1676-1677.

operating as autonomous black boxes, lack both intentions in the dispositional sense and volitions in the immediate causal sense. They process data and optimize outputs algorithmically, without the internal mental-causal experience of willing or choosing. Moreover, Audi’s discussion of deviant causal chains i. e. , outcomes occurring independently of proper volitional control provides a powerful analogy for AI behavior: even when an AI’s conduct produces an outcome that matches a corporation’s goal, the internal process is often non-transparent, stochastic, and causally deviant relative to traditional conceptions of agency. Therefore, attempts to apply human-centered concepts of intent, knowledge, or will to AI actors risk conceptual distortion.

This supports the argument that corporate criminal liability for AIdriven harm must shift focus away from imputing intent to machine agents and instead assess systemic governance failures, compliance structures, and corporate fault at the organizational level. Ohlin’s comparative analysis shows that conceptions of intent differ substantially across legal systems. 86 Ohlin distinguishes between direct intent (dolus directus), knowledge-based intent (dolus indirectus), and risk-based or eventual intent (dolus eventualis). 87 Particularly, dolus eventualis, where an agent foresees the possibility of harm but proceeds regardless, has gained prominence in many civil law jurisdictions and increasingly influences common law thinking in fields such as international criminal law.

88 For example, if a corporation releases an AIdriven trading algorithm that autonomously manipulates financial markets in ways unintended by human developers, but where some degree of risk was foreseeable, the corporation’s culpability could be analogized to dolus eventualis. Thus, the framework for evaluating intent might not require an AI or its operators to have directed a specific outcome purposefully; rather, knowledge of foreseeable risk coupled with indifference may suffice to ground criminal liability under an expanded intent model. This poses an additional challenge to the regulation of AI-related corporate crimes at a global level. Jurisdictions embracing broader notions of intent (e. g. , Germany, France) may be more willing to impose criminal liability for AI-induced harms, whereas stricter mens rea jurisdictions (e. g.

, United States, United Kingdom) may resist unless clear proof of knowledge or purpose is demonstrated. 89 Harmonization of standards thus requires an acknowledgment that the mental element in AI liability is not universally

86 See Jens D. Ohlin, Targeting and the Concept of Intent, 35 Mɪᴄʜ. J. Iɴᴛ'ʟ L.79 (2013)

87 See id at 86–87.

88 See id at 88-100.

89 See id at 87-91.

understood, and any international or cross-border regulatory framework must address these cultural divergences. Recent empirical research challenges the assumption that mens rea must be rooted in natural persons. Stuart and Kneer's study reveals that people readily attribute inculpating mental states; such as knowledge and intention to AI systems, and often assign moral blame on par with or even exceeding that of human agents or corporations. 90 These findings problematize traditional identification and vicarious liability models, which presuppose a natural human ‘mind and will’. Instead, they support a shift towards corporate fault and systems-based models of liability that can accommodate emergent, distributed, and autonomous decision-making by AI agents.

Notably, their identification of an “inverse outcome effect” whereby AI systems are blamed less when harm occurs; illustrates the public’s intuitive awareness of responsibility gaps and the need for a legal framework that does not misallocate blame or excuse culpable developers and corporations. 91 Any legal attribution of mens rea to AI must be carefully structured to avoid undermining human accountability. b. Problem of strict liability and transparency mandates One of the leading responses to the difficulties AI introduces into corporate criminal liability, particularly the collapse of traditional mens rea frameworks, is the proposal to impose strict liability regimes and mandate transparency obligations.

92 Scholars argue that if intentionality, knowledge, or foreseeability cannot be reliably attributed to corporations operating AI systems, it becomes necessary to shift away from subjective fault models and focus instead on objective outcome-based responsibility. 93 90 See Michael T. Stuart & Markus Kneer, Guilty artificial minds: Folk attributions of mens rea and culpability to artificially intelligent agents, Pʀᴏᴄᴇᴇᴅɪɴɢs ᴏғ ᴛʜᴇ ACM ᴏɴ Hᴜᴍᴀɴ-Cᴏᴍᴘᴜᴛᴇʀ Iɴᴛᴇʀᴀᴄᴛɪᴏɴ 5 CSCW2, 1, 1-27 (2021).

91 See id at 12.

92 See Yavar Bathaee, The artificial intelligence black box and the failure of intent and causation, Hᴀʀᴠ. JL & Tᴇᴄʜ 31, 889, 928-931 (2017).

93 See Nora Osmani, The complexity of criminal liability of AI systems, Mᴀsᴀʀʏᴋ Uɴɪᴠᴇʀsɪᴛʏ Jᴏᴜʀɴᴀʟ ᴏғ Lᴀᴡ ᴀɴᴅ Tᴇᴄʜɴᴏʟᴏɢʏ14.1, 53, 53-82 (2020); Christiane Wendehorst, Strict liability for AI and other emerging technologies, Jᴏᴜʀɴᴀʟ ᴏғ Eᴜʀᴏᴘᴇᴀɴ Tᴏʀᴛ Lᴀᴡ 11.2, 150, 150-180 (2020); Sadaf Fahim & G. S. Bajpai, AI and Criminal Liability, Iɴᴅɪᴀɴ J. Aʀᴛɪғɪᴄɪᴀʟ Iɴᴛᴇʟ. & L 1, 64 (2020).

Strict liability frameworks eliminate the requirement of proving mental elements such as intent or recklessness. 94 Instead, liability attaches solely upon the occurrence of a prohibited outcome, provided the defendant engaged in the proscribed act or omission. 95 Under such an approach, corporations deploying autonomous or semi-autonomous AI systems could be held criminally liable for harms caused by those systems without needing to demonstrate specific human intentionality behind each action. Simultaneously, scholars and policymakers advocate for transparency mandates, particularly when corporations use opaque “black-box” AI systems. 96 These mandates would require companies to maintain sufficient documentation of AI design, training data, decision-making parameters, and post-deployment monitoring mechanisms.

Transparency obligations serve two purposes: first, they impose proactive governance duties that encourage safe system deployment; and second, they reduce information asymmetries that obstruct enforcement agencies and courts from scrutinizing harmful AI behavior. 97 c. Problem of causation However, even if strict liability and transparency mandates address the mens rea problem, a second fundamental difficulty remains: establishing causation (actus reus). 98 Criminal liability requires not merely proof that harm occurred but also that the defendant's conduct was voluntary99 and caused the harm in a legally relevant way. 100 There is a need to establish factual 94 See Richard A. Wasserstrom, Strict liability in the criminal law, Sᴛᴀɴ. L. Rᴇᴠ. 12, 731 (1959); David Prendergast, The Constitutionality of Strict Liability in Criminal Law, Dublin ULJ 33, 285 (2011); Kenneth W.

Simons, When is strict criminal liability just, J. Cʀɪᴍ. L. & Cʀɪᴍɪɴᴏʟᴏɢʏ 87, 1075 (1996).

95 See id.

96 See Yavar Bathaee, The artificial intelligence black box and the failure of intent and causation, Hᴀʀᴠ. JL & Tᴇᴄʜ 31, 889, 928-931 (2017).

97 See Heike Felzmann et al. , Transparency you can trust: Transparency requirements for artificial intelligence between legal norms and contextual concerns, Bɪɢ Dᴀᴛᴀ & Sᴏᴄɪᴇᴛʏ 6. 1 (2019); , H. Felzmann, et al. , Towards Transparency by Design for Artificial Intelligence, Sᴄɪ Eɴɢ Eᴛʜɪᴄs 26, 3333, 3333–3361 (2020); Stefan Larsson & Fredrik Heintz, Transparency in artificial intelligence, Iɴᴛᴇʀɴᴇᴛ ᴘᴏʟɪᴄʏ ʀᴇᴠɪᴇᴡ 9. 2, 1, 1-16. (2020); Thomas Wischmeyer, Artificial intelligence and transparency: opening the black box, In: T. Wɪsᴄʜᴍᴇʏᴇʀ & T. Rᴀᴅᴇᴍᴀᴄʜᴇʀ(ᴇᴅs), Rᴇɢᴜʟᴀᴛɪɴɢ Aʀᴛɪғɪᴄɪᴀʟ Iɴᴛᴇʟʟɪɢᴇɴᴄᴇ, 75, 75-101 (Springer International Publishing)(2019). 98 See Arshdeep Ghuman, Elements of crime, Int'l JL Mgmt. & Human. 1, 70 (2018); N. Cross, Criminal justice, actus reus and mens rea, In: Kᴀʀᴇɴ Cᴏʀᴛᴇᴇɴ ᴇᴛ ᴀʟ.

(Eᴅs), Fᴏʀᴇɴsɪᴄ Psʏᴄʜᴏʟᴏɢʏ, Cʀɪᴍᴇ ᴀɴᴅ Pᴏʟɪᴄɪɴɢ, 108, 108– 113 Bristol university online )(2023); A. P. Simester, Causation in (criminal) law, Law Quarterly Review 113, 416, 416-441 (2017); John E. Stannard, Stretching out the actus reus, Iʀɪsʜ Jᴜʀɪsᴛ 1966 28, 200, 200-220 (1993). 99 See R. v. Bourne, [1952] 36 Cr. App. R. 1251 (U. K. ); R. v. Quick, [1973] Q. B. 910 (U. K. ); R. v. Parks, [1992]

2 S.C.R. 871 (Can.).

100 See Paul K. Ryu, Causation in criminal law, U. Pᴀ. L. Rᴇᴠ.106, 773 (1957); Marcelo Ferrante, Causation in Criminal Responsibility, Nᴇᴡ Cʀɪᴍɪɴᴀʟ Lᴀᴡ Rᴇᴠɪᴇᴡ 11.3, 470, 470-497 (2008); Zvonimir Tomic, About Causation in Criminal Law, YBL Fᴀᴄ. Sᴀʀᴀᴊᴇᴠᴏ 49, 459 (2006); Igor Martinovic, The Problem of Causation in Criminal Law, Cʀᴏᴀᴛɪᴀɴ Aɴɴ. Cʀɪᴍ. L. & Pʀᴀᴄ. 19, 75 (2012).

causation and legal causation. Factual causation is to be established via “But for” test i. e. , Would the result have occurred but for the accused’s conduct? 101 While legal causation asks if the defendant’s act is the substantial, operative, and foreseeable cause of the result? 102 Nevertheless, the black box problem renders it exceedingly challenging to ascertain whether the harm would not have transpired “but for” any particular corporate action, such as deployment decisions, training data selections, or system configurations, due to the obscurity of the AI's internal mechanisms that lead to harmful decisions. AI systems, especially those utilising machine learning, can adapt, self-optimize, and change in unforeseen manners throughout their design or deployment phases.

In such instances, although a company's choice to implement the system may be seen as a but-for cause in a general sense, identifying individual business actions that factually resulted in the outcome is precarious. In AI-driven organisations, various stakeholders i. e. , engineers, developers, trainers, and executives provide distinct contributions to the final system. No individual human decision may meet the "but for" criterion when evaluated in isolation, and courts may be reluctant to combine disparate actions from several persons without explicit doctrinal justification. The autonomous functioning of an AI system may be regarded as an intervening act (novus actus interveniens), disrupting the causal connection between the actions of the corporate defendant and the resultant harm.

If the AI modifies its behaviour post-deployment in ways that the business could not have precisely anticipated or regulated, courts may determine that the AI's autonomous decision-making represents a novel, independent cause. Legal causation conventionally necessitates that the harm be a predictable result of the defendant’s actions. In the context of AI, detrimental effects may arise from intricate interactions among the AI's training data, real-world inputs, and operational context. Anticipating all potential harms arising from an AI's operation is frequently unfeasible, engendering significant uncertainty over the foreseeability of such harm at the time of implementation. Consequently, while strict liability and transparency mandates offer promising pathways to sidestep the mental element dilemma, they cannot, by themselves, resolve the deeper difficulty of proving causal links.

101 See B. P. Karnaukh, Causation in Tort Law: Review of the but for Test, Pʀᴏʙs. Lᴇɢᴀʟɪᴛʏ 147, 75 (2019); Pratik Tayal, A New World of Causation in Safeguards: Application of the ‘But for’Test’, Global Trade and Customs Journal 10. 10 (2015). 102 See Antje du Bois-Pedain, Novus actus and beyond: attributing causal responsibility in the criminal courts, The Cambridge Law Journal 80. S1, S61-S90(2021); Grant Firkins, Rethinking causation in english criminal law, Tʜᴇ Jᴏᴜʀɴᴀʟ ᴏғ Cʀɪᴍɪɴᴀʟ Lᴀᴡ 87. 1, 18, 18-38 (2023); R. v. White, (1910) 2 K. B. 124 (U. K. ); R. v. Smith, (1959) 2 Q. B. 35 (U. K. ); R. v. Jordan, (1956) 40 Cr. App. R. 152 (U. K. ); R. v. Michael, (1840) 9 C. & P. 356 (U. K. ); R. v. Pagett, (1983) 76 Cr. App. R. 279 (U. K. ); R. v. Kennedy (No. 2), [2007] UKHL 38 (U. K. ).

d. Mapping Inadequacy: Vicarious, Identification, Aggregate & Culture Models Put simply: when AI performs the act, the human-agent hooks of vicarious and identification liability come loose; when we pivot to systems, aggregate and culture models can sweep too broadly without firm limiting principles or AI-specific benchmarks. The table that follows tracks these shortcomings doctrine by doctrine. Theory How it attributes fault AI-related shortcomings (why it misfits) Net risk Vicarious liability (respondeat superior) Imputes employee/agent’s crime within scope & for corporate benefit No “human agent” at the decision point; emergent model behavior strains scope/benefit tests; vendor/outsourced models blur agency; risks turning deployment of

AI

into constructive agency. Over-inclusion (de facto strict liability) or underinclusion (no agent to impute) depending on framing Identificatio n (alter ego) Attributes mens rea of the “directing mind/senior manager” to the company Hard to locate a single culpable human mind when harm flows from system design, training data, or drift; AI outputs lack human mens rea; decisions are diffuse across teams/providers. Under-inclusion (no senior mind to pin)

Corporate fault / culture model Fault inferred from policies, incentives, control systems Black-box opacity frustrates assessing “reasonable” controls; static programs vs. dynamic models (retraining/drift); weak leverage over third-party/vended

AI;

vague benchmarks invite paper compliance. Ambiguous (can be vacuous or overly elastic) IV. Evaluating Indian Legal Frameworks for Corporate Liability

Section 1 (26) of Bhartiya Nyaya Sanhita 2023103 (previously section 11 Indian Penal Code, 1860)104 defines “person” includes any company or association or body of persons, whether incorporated or not. The word 'person' has also been defined in general clauses act under section 3(42) which is similar to that of Bhartiya Nyaya Sanhita 2023. 105 At first, it was believed that a corporation could be held accountable for just those actions for which no mens rea was needed and merely a fine may be assessed. 106 Therefore, a company could not be charged with crimes for which only incarceration is a punishment. In MCD vs. J. B. Bottling Co. (P) Ltd. , a case involving punishment of imprisonment and fine, the court held that the company did not enjoy impunity from prosecution and will be punished with fine only. 107 While in Assistant commr. Assessment vs. Velliappa Textiles Ltd.

,108 the Supreme Court of India held that offences which prescribe a mandatory sentence of imprisonment as punishment, punishing a corporation with fine only, would not be possible without making legislative changes. In the case of Standard Chartered Bank v. Directorate of Enforcement109 a Constitution Bench held that a company can be prosecuted and convicted for an offence which requires a minimum sentence of imprisonment. However it was observed that,

103 See Bharatiya Nyaya Sanhita, 2023, § 1 (22) No. 45, Acts of Parliament, 2023 (India);

104 See Indian Penal Code, 1860, § No. 45, Acts of Parliament, 1860 (India).

105 See Satish Jayantilal Shah v. Pankaj Mashruwala, (1997) BC 320; I (1997) CCR 603; 1996 Cri. L.J. 3099 (India). 106 See Punjab Nat’l Bank v. A.R. Gonsalves, 1921 SCC OnLine Sind JC 31 (India); Ananth Bandu v. Corp. of Calcutta, 1952 SCC OnLine Cal 122 (India).

107 See MCD v. J.B. Bottling Co. (P) Ltd., 1975 SCC OnLine Del 47 (India).

108 See Asst. Comm’r of Assessment v. Velliappa Textiles Ltd., (2003) 11 SCC 405 (India).

109 See Standard Chartered Bank v. Directorate of Enf’t, (2005) Supp. (1) S.C.R. 49 (India).

“8. It is only in a case requiring mens rea, a question arises whether a corporation could be attributed with requisite mens rea to prove the guilt. But as we are not concerned with this question in these proceedings, we do not express any opinion on that issue. ” The Supreme Court in Iridium India Telecom Ltd. v. Motorola Inc. 110 addressed whether a corporation can be held criminally liable for offences requiring mens rea. Drawing from English and American jurisprudence, the Court affirmed that corporations, though artificial entities, can possess a culpable mental state through the doctrine of attribution. 111 This doctrine imputes the mental state of those who constitute the “directing mind and will”—typically directors and top management to the corporation itself. The Court cited Director of Public Prosecutions v. Kent and Sussex Contractors Ltd. 112 and Bolton (Engineering) Co.

Ltd. v. T. J. Graham & Sons Ltd. ,113 where it was held that companies act through their controlling individuals, and their intentions can be treated as the company’s own. Similarly, in Tesco Supermarkets Ltd. v. Nattrass,114 it was clarified that when such individuals act within their sphere of authority, their guilty mind is attributed to the company, not vicariously but as the company’s own mens rea. Consequently, the Court concluded that corporations are liable for both statutory and common law offences involving mental fault, provided those directing the company's affairs possess the requisite mens rea at the time of the offence. While Iridium India Telecom Ltd. v. Motorola Inc.

115 affirmed that a corporation can be held criminally liable for offences involving mens rea through the doctrine of identification/alter ego, imputing the intent of the directing mind to the company, it did not settle the question of automatically extending such liability to individual directors. This distinction was clarified in Sunil Bharti Mittal vs. Central Bureau of Investigation116, wherein the Supreme Court reiterated that individual liability cannot be presumed merely from one’s position in the corporate hierarchy. In paras 37 to 39 of the judgment, the Court underscored the cardinal principle of criminal law that vicarious liability must be explicitly

110 See Iridium India Telecom Ltd. v. Motorola Inc., (2011) 1 SCC 74 (India).

111 See id at Paragraph 59-61.

112 See Dir. of Pub. Prosecutions v. Kent & Sussex Contractors Ltd. & Bolton (Eng’g) Co. Ltd., [1944] K.B. 146

(U.K.)

113 See Bolton (Eng’g) Co. v. T.J. Graham & Sons Ltd., [1957] 1 Q.B. 159 (U.K.).

114 See Tesco Supermarkets Ltd. v. Nattrass, [1971] 2 All E.R. 127 (H.L.) (U.K.).

115 See Iridium India Telecom Ltd. v. Motorola Inc., (2011) 1 SCC 74 (India).

116 See Sunil Bharti Mittal v. Cent. Bureau of Investigation, (2015) 1 S.C.R. 377 (India).

provided by statute. Thus, an individual may only be held liable if (i) there is clear evidence of their active role coupled with criminal intent, or (ii) a statute specifically incorporates a vicarious liability clause, such as Section 141 of the Negotiable Instruments Act, 1881. Consequently, the jurisprudence draws a clear line between corporate fault and personal culpability, requiring distinct thresholds for attributing liability to natural and juristic persons. In Shiv Kumar Jatia v. State (NCT of Delhi), the Supreme Court reaffirmed the principle that in the absence of statutory vicarious liability, individual directors of a company can only be held criminally liable if there is clear evidence of their active role coupled with criminal intent. 117 The Court quashed the criminal proceedings against the managing director (MD) of Asian Hotels (North) Ltd.

, operator of Hotel Hyatt Regency, New Delhi, where a guest fell from an unsecured terrace. While negligence was alleged against hotel staff, the MD was charged merely due to his position and signature on official documents, without any direct attribution of culpable intent. The Court relied on its earlier ruling in Sunil Bharti Mittal v. CBI,118 where it had held that criminal liability of directors requires more than mere designation; it must be supported by substantive material showing personal involvement and mens rea. The Court also referred to Maksud Saiyed v. State of Gujarat,119 emphasizing that the Indian Penal Code (IPC) does not provide for vicarious liability, and any imposition of liability must meet the threshold of specific allegations indicating individual culpability. Vague or generalized assertions, as in the present case, were deemed insufficient.

Accordingly, the Court quashed the proceedings against the MD but allowed trial to continue against the general manager, noting that his role in the day-to-day operations warranted factual inquiry at trial, regardless of his claimed absence from the premises on the date of the incident. Similarly, in S. M. S. Pharmaceuticals Ltd vs Neeta Bhalla And Anr. ,120 observed that, “8 .... There is no universal rule that a Director of a company is in charge of its everyday affairs. We have discussed about the position of a Director in a company in order to illustrate the point that there is no magic as such in a particular word, be it Director, manager or secretary. It all depends upon the respective roles assigned to the officers in a company. A company may have managers or secretaries for different departments,

117 See Shiv Kumar Jatia v. State (NCT of Delhi), AIR 2019 SC 4463 (India).

118 See Sunil Bharti Mittal v. Cent. Bureau of Investigation, (2015) 1 S.C.R. 377 (India).

119 See Maksud Saiyed v. State of Gujarat, AIR 2007 SC 332 (India).

120See S.M.S. Pharms. Ltd. v. Neeta Bhalla, (2005) 8 SCC 89 (India).

which means it may have more than one manager or secretary. "This judgment reinforces settled jurisprudence that individual liability under the IPC, absent statutory vicarious provisions, arises only upon proof of personal actus reus and mens rea. Directors cannot be held criminally liable solely based on their official designation or participation in board meetings. ” While in Aneeta Hada v. Godfather Travels & ·Tours (P) Ltd. 121 5 (2012), it was observed: “32. We have referred to the aforesaid authorities to highlight that the company can have criminal liability and further, if a group of persons that guide the business of the companies have the criminal intent, that would be imputed to the body corporate. In this backdrop, Section 141 of the Act has to be understood.

The said provision clearly stipulates that when a person which is a company commits an offence, then certain categories of persons in charge as well as the company would be deemed to be liable for the offences under Section 138. Thus, the statutory intendment is absolutely plain. As is perceptible, the provision makes the functionaries and the companies to be liable and that is by deeming fiction. A deeming fiction has its own signification. ” In S. K. Alagh v. State of U. P122 also it was observed that “19.... In absence of any provision laid down under the statute, a Director of a Company or an employee cannot be held to be vicariously liable for any offence committed by the Company itself. ” These judgments reinforce the settled jurisprudence that individual liability under the IPC/BNS, absent statutory vicarious provisions, arises only upon proof of personal actus reus and mens rea.

Anyone cannot be held criminally liable solely based on their official designation. Broadly, two categories of liability emerge under Indian law:

1. Personal criminal liability: Where no vicarious liability clause exists, liability can only

be imposed upon proof of direct involvement in the offence with requisite intent.

121 See Aneeta Hada v. Godfather Travels & Tours (P) Ltd., (2012) 5 SCC 661 (India).

122 See S.K. Alagh v. State of U.P., (2008) 5 SCC 662 (India).

2. Statutory vicarious liability: Various statutes such as the Companies Act, 2013, impose

vicarious liability on designated persons, including “officers in default,” directors responsible for particular functions, and those acting with consent or connivance. a. Analysing penal provisions of Companies Act, 2013 In light of the Supreme Court’s consistent emphasis on the necessity of express statutory provisions for attributing vicarious criminal liability, the framework under the Companies Act,

2013 becomes especially significant. Unlike the Indian Penal Code (now Bhartiya Nyaya

Sanhita), which lacks a general doctrine of vicarious liability, the Companies Act adopts a more structured approach by statutorily identifying individuals. The Companies Act, 2013 serves as a fundamental framework for corporate regulation in India.123 It establishes a thorough compliance framework for corporate entities and their officials to safeguard the rights and interests of shareholders and investors.124 The incremental decriminalisation of procedural defaults under the Act, as advised by the Injeti Srinivas Committee, underscores the legislative purpose to reconcile severe enforcement with the necessity of fostering a business-friendly regulatory framework.125 The Act includes an internal adjudication process, as specified in Section 454 of the Act in conjunction with the Companies (Adjudication of Penalties) Rules,

2014.126 The MCA, by its notification of March 24, 2015, designated the RoC of the respective

jurisdictions to serve as adjudicating officers. 127 This internal adjudication system, however, is applicable solely to those non-compliances or violations under the Act that incur only civil liability, specifically a monetary penalty. 123 See Sarah Alvy, India's Companies Act of 2013: A Governance Shift into the Sunlight, Iɴᴅᴏɴ. J. Iɴᴛ'ʟ & Cᴏᴍᴘ. L. 2, 187 (2015); Harshit Chowdhary, Corporate Governance Practices in India: A Theoretical Aspect, LᴀᴡFᴏʏᴇʀ Iɴᴛ'ʟ J. Dᴏᴄᴛʀɪɴᴀʟ Lᴇɢᴀʟ Rsᴄʜ. 2, 173 (2024); Aʀᴜɴᴀᴄʜᴀʟᴀ Rᴀᴍᴀɪʏᴀ, Gᴜɪᴅᴇ ᴛᴏ ᴛʜᴇ Cᴏᴍᴘᴀɴɪᴇs Aᴄᴛ (LexisNexis)(2021). 124 See Uma Rani, Shareholder Rights and Corporate Governance: An Analysis of Recent Developments in India, Issue 5 Iɴᴛ'ʟ JL Mɢᴍᴛ. & Hᴜᴍᴀɴ.

7, 1452 (2024); Mihir Naniwadekar & Umakanth Varottil, The stakeholder approach towards directors’ duties under Indian Company Law: a comparative analysis In: Mᴀʜᴇɴᴅʀᴀ Pᴀʟ Sɪɴɢʜ, Tʜᴇ Iɴᴅɪᴀɴ Yᴇᴀʀʙᴏᴏᴋ ᴏғ Cᴏᴍᴘᴀʀᴀᴛɪᴠᴇ Lᴀᴡ, 95-120 (Oxford University Press)(2017); Neeti Shikha, Corporate governance in India-the paradigm shift, Iɴᴛᴇʀɴᴀᴛɪᴏɴᴀʟ Jᴏᴜʀɴᴀʟ ᴏғ Cᴏʀᴘᴏʀᴀᴛᴇ Gᴏᴠᴇʀɴᴀɴᴄᴇ 8. 2, 81, 81-105 (2017); Devarshi Mukhopadhyay & Rudresh Mandal. The end of shareholder primacy in Indian corporate governance? Says who? , Cᴏᴍᴍᴏɴᴡᴇᴀʟᴛʜ Lᴀᴡ Bᴜʟʟᴇᴛɪɴ 46. 4, 595, 595-610 (2020). 125 See Injeti Shrinivas et al. , Report of the committee to review offences under companies act, 2013, Mɪɴɪᴛsᴛʀʏ ᴏғ ᴄᴏʀᴘᴏʀᴀᴛᴇ ᴀғғᴀɪʀs (Aug. 2018) (https://www. mca. gov. in/Ministry/pdf/ReportCommittee_28082018. pdf) (Last accessed on May. 5, 2025) 126 See Companies Act, 2013, No.

18 of 2013, § 454, Act of Parliament (India); MCA, Companies (Adjudication of Penalties) Rules, 2014, New Delhi, G. S. R. 254 (E), Mar. 31, 2014 (https://ibclaw. in/the-companies-adjudication-of-penalties-rules-2014/) (last accessed May 4, 2025).

127 See id Rule 3.

The Companies Act, 2013 also permits compounding of certain offences, those not punishable with imprisonment only or with imprisonment and fine by allowing companies and their officers to admit non-compliance and pay a prescribed fee as determined by the Regional Director or NCLT. 128 This mechanism, governed by Section 441, enables rectification of defaults without prolonged litigation. While no express preconditions exist, defaults must typically be remedied before applying. Companies and officers may file joint applications, even for repeated offences over multiple years. The adjudicating authority enjoys broad discretion in determining the compounding amount, considering factors such as intent, repetition, public interest, and financial condition.

However, the procedure faces challenges, including absence of statutory timelines, interpretational inconsistencies, and a rigid bar on compounding repeat offences within three years, which may unfairly penalize companies despite genuine efforts to comply. The analysis of penal provisions of Companies act, 2013 reveals that: (i) Certain provisions target only specific individual roles, like managing directors, whole-time directors, CFOs, or any specifically designated individuals wherein certain duty is imposed on these personnel and they contravene such duty. For instance, Section 128(6),129 129(7),130 166(7),131 and others. These reflect a strict liability approach.

128 See Companies Act, 2013, No. 18 of 2013, § 441, Act of Parliament (India);

129 See Companies Act, 2013, No. 18 of 2013, § 128(6), Act of Parliament (India). [§128 (6) If the managing director, the whole-time director in charge of