Unveiling Insights on Legal Approaches to AI Liability: From Code to Courtroom

I. Introduction

In India, artificial intelligence (AI) is transforming a number of sectors, including the legal system. Artificial intelligence is becoming a commonplace aspect of our daily existence. In one out of every three decisions that impact our personal and professional destinies, artificial intelligence (AI) is present, from the algorithms that suggest Netflix shows to the systems that employers use to assess job applicants. Though AI may seem to offer impartiality and efficiency, it frequently reflects the prejudices of the society that developed it. AI is far from neutral; in fact, it can feed prejudices and societal injustices, particularly when it comes to racial, gender, and cultural stereotypes. Additionally, an AI system is capable of making decisions via inference engines and acting in the real world. The capacity to grow via experience and to adjust over time enhances these talents.In 2024, the majority of AI researchers and practitioners³, as well as the majority of AI-related headlines, are focused on breaking through. AI-enabled applications and devices can see and identify objects, understand and respond to human language, learn from new information and experience, make detailed recommendations to users and experts, and act autonomously, eliminating the need for human intelligence or intervention (a self-driving car is a classic example).In order to grasp generative AI completely, it is crucial to comprehend the foundational technologies like machine learning (ML) and deep learning that underpin generative AI tools. As AI technology becomes more widespread, concerns about its potential effects on criminal responsibility emerge.

II. Background

The research that underlies AI's criminal responsibility encompasses several legal, ethical, and technological components. Initially, talks on AI and culpability focused on traditional legal theories like negligence and strict liability, but these frameworks could not adequately reflect the complicated cities brought by autonomous systems.Early incidents of AI, such as automated vehicles and software malfunctions, highlighted concerns about responsibility. However, these cases frequently center on human operators or producers rather than AI itself.Many jurisdictions are now studying legislation relevant to Ai. The European Union's AI Act is a notable example that seeks to regulate high-risk AI applications and explain liability issues. Current research looks into several liability models, including holding developers, producers, and users liable for AI actions.The concept of product liability is frequently explored in the context of autonomous systems. As AI technology advances, there will certainly be requests for complete legal reforms to handle the intricacies of AI liability, potentially resulting in the creation of new types of liabilities particular to AI.

(A) Literature Review

The author emphasizes the critical necessity for the Indian legal system to handle the complicated consequences of AI's role in criminal responsibility. To ensure the impartiality and integrity of the criminal justice system, efforts should focus on enforcement and sentencing. In India, loss due to AI should include the concept of strict liability. This means that creators and operators would be held liable for the activities of AI regardless of their goals or understanding (Hifatajali Sayyed,2024)According to the author, under any country's state legislation, the criminal accountability of AI robots is unclear. As a result, only court statements are used as a primary source of judgment in cases where AI commits a specific offense. The liability model is only relevant if legislators understand how to amend existing laws or create new ones to address the issue of AI system accountability as its impact on human life expands(Thanush).The author suggests a legal system that holds software owners, developers, or creators liable for actions or inactions that could have been predicted due to AI flaws. The author suggests providing AI-powered software and robots legal personas, including rights and obligations, to prevent any issues.However, as AI is still growing as a technology, it is not projected that such a sophisticated system will arise in India in the near future(Antara Roy,2024).The study examines three features of the EU liability framework for AI systems. 1. Rules governing the scope of responsibility 2. The accountable parties. 3. Standard of Liability. The European Commission proposes horizontal responsibility standards for producers and AI-specific restrictions for users and system owners. It focuses on assigning responsibility for AI-related harm to AI makers and users. If politicians and courts correctly define the duty of care, the owner is compelled to take effective precautions(Miriam Buiten,2023).This article explores how governments view the liability of AI. Many nations struggle with determining who is responsible for AI. Some believe the company that created the AI is liable, while others believe the person who created the algorithm is. Others believe the company that purchased and used the AI software is responsible.The author employs the natural likely consequence liability model, which holds programmers and users accountable for AI-related crimes, even if they had no intention of doing them(Shyamal Dave,2023).

(B) Objectives

• To study the relationship between artificial Intelligence and its liability.
• To analyze the existence and impact of Artificial Intelligence in India.
• To suggest General ways regarding criminal liability of Artificial Intelligence.

(C) Research Problem

In this paper, the author attempts to address artificial intelligence's participation in criminal activities in India.The author also answers to the question of whether artificial intelligence should be given legal standing under Indian legislation.The author also discusses the development of a legal framework that includes vicarious liability, product liability, and strict liability for AI systems, successfully balancing the responsibilities of producers and users while providing equitable recompense for AI-related harm.

(D) Research Questions

1. What are the implications of granting legal personhood to Artificial Intelligence ?

2. How can a clear duty of care be defined for both producers and users of artificial intelligence ?

(E) Research Method

This article utilizes doctrinal research methodology, a tool used in legal research, to explore and support a thorough understanding of the topic. The author uses secondary data sources to guide the research objectives and inquiries.The analysis of the study is descriptive in nature. It seeks to identify numerous aspects and qualities.Several aspects of the research subject were investigated. Data for this study were gathered from many sources, including government publications, censuses, internal organizational records, journal articles, websites, reports, and books.

III. Ai regulations in india

Currently, AI is not particularly governed by any written laws, legal regulations, or guidelines in India. As to The Indian Express, the Ministry of Electronics and IT (MeitY)⁴ is drafting a new law that will offer rules and advice to encourage the responsible advancement and usage of AI technologies, acknowledging its benefits, instead of penalizing those who violate them. Nevertheless, a number of mechanisms are being created to supervise AI regulations.

• National Artificial Intelligence Strategy: In 2018, Niti Ayog announced the inaugural national AI plan, #AIFORALL, aimed at having an all-encompassing approach to artificial intelligence. The plan identified key industries, including healthcare, education, agriculture, smart cities, and transportation, for national emphasis in AI advancement and application. Some of the strategy's recommendations have now been put in place, including creating excellent quality data sets to support research and development and creating legislative⁵ frameworks to protect cybersecurity and data.
• DPDP Act : On August 11, 2023, the President of India formally enacted the Digital Personal Data Protection Act. This law, which goes into force right away, regulates how digital personal information is handled in India, regardless of how it was collected. It also addresses some privacy issues related to artificial intelligence (AI) technologies.
• Information Technology (Intermediary Guidelines and Digital Media Ethics Code), 2021: The IT Rules 2021, mandated by the Government of India according to the Information Technology Act of 2000, give a structure to regulate social media intermediaries, OTT platforms, and digital news media. The rules were updated on April 6, 2023, and went into force on May 26, 2021.
• Information Technology Act, 2000 : The Information Technology Act 2000 (IT Act) and the Information Technology (Reasonable security practices and procedures and sensitive personal data or information) Rules 2011 are two laws that have an impact on AI in India. The primary legislation in India pertaining to cybercrime and electronic trade is the IT Act. It also addresses the laws pertaining to artificial intelligence.

IV. Benefits of legislations holding ai accountable for its actions

• Encourages innovation: Companies might be more inclined to invest in and develop AI technology and nurture discoveries if the liability of creators is limited.
• Consumer protection: Consumer confidence is increased when systems are held responsible for their activities, as this guarantees users have options in the event of damage or breakdown.
• Clear legal framework: By establishing liability particular to AI, a clear legal framework is created that can facilitate regulatory and dispute resolution procedures.
• Risk management: By understanding that the technology bears responsibility instead of their personnel or establishments, it enables businesses to more effectively manage the risk involved in implementing AI.
• Focus on AI governance: Governments can refocus their efforts on developing policies and governance frameworks that guarantee the moral application of AI by making AI accountable.
• Promote Responsibility in design: If developers are aware that AI may be held accountable, then they might be more likely to address ethical and safety concerns in their designs.
• Facilitates trust in AI: If people believe that AI systems are accountable for their activities it can generate trust in technology leading to greater adoption and utilization.
• Encourages better regulation: It might necessitate the creation of an improved legal structure specifically designed to address AI, guaranteeing the safe operation of these systems in society.

V. Regulation of ai around the world

In response to its explosive expansion, a few nations, like Japan and Israel, have clarified data, privacy, and copyright laws, paving the path for the use of copyrighted content in AI training.

• Brazil - After three years of proposed (and blocked) measures⁶ On the topic, Brazil now has a draft AI law.Because of the law's emphasis on users' rights, AI suppliers are required to tell users about their products. Users are entitled to both knowledge that they are dealing with artificial intelligence (AI) and an explanation of the reasoning behind any recommendations or decisions the AI made. All AI developers are liable for damage caused by their AI systems, while developers of high-risk goods are held to an even higher threshold reached of culpability.
• China - China is looking for feedback from the public on a draft regulation for generative AI. China's proposal, however, states that generative AI must adhere to "Socialist Core Values," unlike most other nations.There are other limitations on where training data can be obtained; developers risk legal repercussions if their training material violates the intellectual property rights of third parties. Additionally, the rule mandates that AI services be created with the sole purpose of producing "true and accurate" material.
• European Union - On August 1st, the new AI law of the European Union went into effect. Most importantly, it establishes standards for various AI systems according to the degree of risk they represent. Stronger requirements⁷ are imposed on AI systems the greater the risk it presents to people's health, safety, or human rights.Australia and other nations may learn a lot from this law, which is the first of its kind in the world to address AI risks completely, as they work to make AI safe and advantageous for all.
• Japan - Similar to Israel, Japan has embraced a "soft law" approach to AI regulation, meaning that there are no strict laws dictating how AI may or cannot be utilized. Instead, claiming a wish to avoid restricting innovation, Japan has chosen to wait and observe how AI evolves.As of right now, Japan's AI developers are forced to follow rules that are closely related to their work, like data protection laws.

VI. Can legal status be given to ai

The main query is whether AI qualifies as a legal entity. Legal entities are those that the law acknowledges as having the capacity to be bound by obligations and rights in legal interactions. Legal entities fall into two categories: artificial and natural individuals. Article 21 of the Indian Constitution defines legal personality as a feature of individual sovereignty. Article 21 states that a person cannot be deprived of their life or freedom unless a legally recognized process has been followed. Legal personhood is not restricted to individuals in India; nonetheless, it has not been extended to technology. Furthermore, the Companies Act creates a legal framework that regards enterprises as distinct legal entities, thus establishing a precedent for the legal recognition of AI.

Companies and AI are not the same; corporations have stakeholders that hold them responsible, whereas AI has a great deal of autonomy. The only country that acknowledges artificial intelligence (AI) as a legitimate entity is Saudi Arabia, which also granted citizenship to Sophia, an AI humanoid, with the same⁸ rights and obligations as human citizens.The evolving legal position of AI agents is one concern raised by its use and advancement. Artificial Intelligence is still in its infancy and lacks legal recognition in India. Current laws do not precisely define artificial intelligence, and the complexity of robots makes it difficult to determine how standard restrictions would apply.

Both 'actus reus' and 'mens rea' are necessary for someone to be found guilty in a criminal case. If an AI satisfies both of these requirements, there's no excuse not to hold it directly responsible for the misconduct. An AI robot satisfies⁹ the actus reus criteria if it physically injures someone with its hydraulic arm. Similarly, an AI entity can be held accountable for its actions if it is assigned a responsibility and does not perform it. The real challenge is in holding AI legally accountable for misconduct. to prove that AI was aware of and intending to commit crimes, or that it possessed mens rea. Humans use their eyes, ears, tongues, noses, and skin to gather information, which is then interpreted by the brain to control how they behave. Similar activities are carried out by advanced artificial intelligence technology.Information is acquired from several sources. gathers data, assesses, analyzes, and decides what to do next. Artificial Intelligence is able to think faster and more effectively than humans. The same criminal rules might apply to AI, although with a few minor adjustments.

VII. Liability of artificial intelligence

In accordance with general torts law, contracts, and statutes as the case may allow, anyone who has experienced loss or harm as a result of AI has a number of possibilities for pursuing claims for compensation against the manufacturer, owner, keeper, user, network provider, software supplier, etc. Supporting these claims is an essential lawful guideline that is codified in the Latin proverb "ubi jus ibi remedium," which translates to "where there is a wrong, there is a remedy."

1. Vicarious liability - The legal doctrine known as "vicarious liability" holds an individual accountable for the deeds of another, not because they committed the wrong themselves, but rather because of their connection to the offender.Some people contend that, depending on the circumstances, AIs should be seen as agents of their manufacturers or owners and that, in such cases, the human principal should be held vicariously accountable for any harm the AI causes. This is justified by the fact that artificial intelligence (AI) is made to carry out tasks assigned by humans and to respond to their commands. Accordingly, it has been proposed that the human principal could be held accountable for the harm the AI agent causes through the application of vicarious liability. The example of robots and autonomous vehicles is one that is frequently used to illustrate this. It has therefore been suggested that the corporation, acting as the "employer" of the robot, should be held vicariously accountable for circumstances where, for example, a robot is programmed to carry out tasks in a firm that would typically be undertaken by a human being¹⁰ (an employee) and damage as a result. Only when an agent is acting in the course of their employment is a principal liable for the agent's activities. Courts will need to decide whether AI systems can fall under a special kind of agency law as they get more flexible and self-learning.

2. Product liability - According to the basic theory of product liability, when a "product" is sold to an end user or consumer in a defective or unreasonable dangerous condition that results in bodily harm to them or their property, the manufacturer, seller, or any other party involved in the product's sale and distribution chain is liable for damages. In the case of AI, product liability becomes a matter needing careful evaluation based on its usage context (both intended and unplanned).In general, three (3) bases may give rise to product liability if an AI-incorporating offering is governed by the applicable law (i.e., is a "product" as defined by statute or case law): manufacturing defects, design defects (which are frequently hard to distinguish from manufacturing defects), and warnings defects. The Nigerian Supreme Court applied the common law principle established in Donoghue v. Stevenson in Nigerian Bottling Company Limited v. Ngonadi, which states that if a manufacturer violates their duty of care and causes harm to a party, the manufacturer may be held accountable for the harm if it was a reasonably foreseeable result of the manufacturer's actions.Such product liability claims have the drawback of just compensating the manufacturer for harm; they do not impose obligation on the AI's owner, keeper, user, network provider, software supplier, etc. Companies that incorporate artificial intelligence (AI) into their product and service offerings must create a comprehensive risk management framework (RMF) with governance policies, procedures, and processes to guard against a variety of possible internal and external AI threats.

3. Strict liability - Instances where a party is held accountable for losses or harm without having to establish negligence or fault are known as strict liability.If an AI system is found to be flawed or unduly risky, strict responsibility may be imposed when the system causes harm. For example, if a self-driving car fails and causes an accident, and it turns out that the vehicle has software problems or design flaws, the manufacturer might be held strictly accountable. AI does not now have legal personhood. This implies that, unlike people or companies, AI cannot be held accountable on its own. Typically, those that create, run, or utilize AI are held accountable. Depending on the way the system was created, tested, and put into use, claims against AI systems may still frequently need to show negligence as opposed to strict responsibility.Strict liability for artificial intelligence is a complicated legal matter that differs depending on the situation and the country.

VIII. Responsibilities of producers and users of artificial intelligence

• When using AI systems, producers need to perform extensive risk evaluations and safety studies.
• Make certain employees utilizing AI systems have received the necessary training to comprehend the technology and how it will affect their jobs.
• Recognize the biases and mistakes that artificial intelligence (AI) systems may have, and try not to rely too much on their results.
• Adhere organizational rules and guidelines for the use of AI, especially those pertaining to data security and moral behavior..
• Manage data sensibly, making sure that any information utilized with AI systems conforms with security and privacy requirements.

IX. Case laws addressing liability of ai

There are currently few specific case laws, especially in India, that deal directly with the liability of AI systems. Nonetheless, a number of pertinent decisions from different jurisdictions offer guidance on how judges should handle questions of AI responsibility. Legal problems about personhood and accountability for actions committed by autonomous systems have been raised by discussions surrounding the liability of Sophia, a humanoid robot that was granted citizenship in Saudi Arabia, even though there isn't a real court case surrounding it. This instance serves as a reminder of how difficult it is to assign blame to AI. Grubbs v. Nissan North America (2016): In this instance, a plaintiff accused Nissan of causing an AI-driven vehicle to crash into an object. Even while the case mostly dealt with traditional product liability, it established a precedent for situations in the future when harm caused by AI systems may be suspected.Gordon v. Google, Inc. (2015): In this case, Google was accused of algorithmic defamation. It looked at who would be responsible and if AI-driven search results may be considered defamatory. The conclusion highlighted the roles that platforms must play in overseeing content created by AI. The debate on autonomous car liability in the UK has been developing. In its consideration of who should be held accountable for self-driving car accidents, the Law Commission of England and Wales has proposed that manufacturers be held accountable in the event that their technology malfunctions.

Although India lacks case law specifically on AI liability, the concepts of carelessness found in tort law may nonetheless be relevant.

X. India and ai liability

Particularly in India, the legal environment¹¹ around AI liability is still evolving. Nonetheless, a number of current frameworks and regulations may be applicable to matters pertaining to AI, particularly in regards to liability. If an AI system causes injury as a result of developer or user negligence, the IPC's negligence rules may be enforced.Under certain conditions, responsibility may be addressed in certain sections when an AI's activities cause criminal harm. AI systems that are employed in cybercrime may be subject to the provisions of this Act, which deals with electronic crimes.It describes how intermediaries—which can include AI service providers—are liable for material produced by their systems.This Act covers AI systems used in consumer items and makes service providers and manufacturers responsible for any faults and resulting harm. These recommendations can assist regulate the liabilities of platforms that use AI-driven algorithms, as AI has an expanding influence on e-commerce. Liability provisions, which specify the extent to which developers and users may be held liable for the AI's activities, can be included in contracts regulating AI services.In the event that an AI does not live up to expectations, parties may file a breach of contract lawsuit. If users and creators of AI fail to guarantee the security and dependability of their systems, then existing tort concepts about negligence may be applicable to them. The laws governing product liability may compel manufacturers to pay damages for flaws or faults in AI products if they cause injury.

(A) Punishments for Artificial Intelligence

India is going to be tough on deep fakes and artificial intelligence misinformation. They will do this with the aid of a new legislation that may allow for the fines of writers and social media platforms that facilitate the dissemination of this sort of destructive information.The government and other stakeholders will create actionable items in ten days on identifying deep fakes, halting their uploading and viral sharing, and bolstering the reporting mechanism for such content, according to a recent statement made by Union Minister of Information Technology and Telecom Ashwini Vaishnaw¹². It will make it possible for consumers to sue websites that host harmful AI-generated content.In the first scenario, it is assumed that the AI entity is an innocent agent carrying out the user's orders. In such a situation, criminal culpability may result from the producer's deliberate programming to conduct an offense or the user's improper usage of the AI entity to accomplish the offense.Furthermore, by compensating the harmed parties, the inventors and developers of artificial intelligence are also held accountable for the harm that their technology has produced.Additionally, artificial intelligence can be disciplined by having its use suspended for a set amount of time.

(B) Benefits of legislations holding ai accountable for its actions

• Encourages innovation: Companies might be more inclined to invest in and develop AI technology and nurture discoveries if the liability of creators is limited.
• Clear legal framework: By establishing liability particular to AI, a clear legal framework is created that can facilitate regulatory and dispute resolution procedures.
• Risk management: By understanding that the technology bears responsibility instead of their personnel or establishments, it enables businesses to more effectively manage the risk involved in implementing AI.
• Focus on AI governance: Governments can refocus their efforts on developing policies and governance frameworks that guarantee the moral application of AI by making AI accountable.
• Promote Responsibility in design: If developers are aware that AI may be held accountable, then they might be more likely to address ethical and safety concerns in their designs.
• Facilitates trust in AI: If people believe that AI systems are accountable for their activities it can generate trust in technology leading to greater adoption and utilization.

(C) Scope

This article explores potential solutions to the problem, including holding artificial intelligence legally accountable for its deeds when it acts autonomously and is uncontrollable. However, punishing AI is unnecessary since it may lead to significant costs and need significant changes to the law. A more efficient way to tackle AI crime is to make small changes to the existing criminal laws that target people and maybe expand civil liability.

(D) Limitations

• This study only focused on the doctrinal viewpoint.
• As a result of the dependence only on secondary sources, the researcher may present a skewed picture of AI.
• As the research was completed in a short time, it was not able to be further expanded.
• By using diverse research approaches, future researchers can expand the scope of their work.

(E) Findings

As AI becomes more sophisticated, autonomous, and complicated, it will become impossible to hold any one person criminally accountable for crimes that AI commits. However as AI is still in its earliest stages, there are currently no general AIs or super AIs, and very few AIs are connected to illegal activity.However, AI is still heavily influenced by humans. Rather than punishing AI, finding other solutions is preferable.The scope of criminal and civil sanctions can be increased as an alternative to penalizing AI directly.

(F) Suggestions

Artificial intelligence needs to be given personhood and be governed by the same criminal laws as companies, which are regarded as legal persons and susceptible to criminal penalties. Therefore, penalizing AI is the best course of action. If AI creators, owners, users, or managers don't take their duties seriously, they might suffer dire repercussions. Investigating AI's participation in criminal culpability in India is a difficult and developing undertaking that might prove to be more fruitful than prosecuting AI directly while penalizing stakeholders indirectly. India may encourage a comprehensive approach that puts ethics, justice, and responsibility first by recognizing the challenges and legal repercussions that AI technology presents. By ensuring AI is incorporated into the criminal justice system in a responsible and ethical manner, dangers may be reduced and its potential can be fully used for the good of society as a whole.For more credibility, the researcher of the future might choose to use primary data. Further information and research are needed to confirm the precise details of AI's criminal culpability in India.

XI. Conclusion

AI focuses on basic logic and lacks the creative thinking that humans are capable of. Robots need to be passionately intelligent, empathetic, and able to follow commands from above. Given that artificial intelligence is still susceptible to risks, integrating technology into the legal sector presents a number of obstacles. Giving artificial intelligence (AI) legal personhood could not solve the worries about its growth and might even create new ones, such putting AI in charge of liabilities and raising the risk of abuse or, in the worst case scenario, exploiting its own powers. Though there aren't many case laws specifically addressing AI liability, there are still plenty of related cases and continuing conversations that point to a changing legal environment. Courts will probably hear more cases involving AI as the technology develops and permeates everyday life, necessitating the creation of more precise legal norms and liability frameworks. Future legal decisions will therefore have a significant influence on how liability is decided in the context of AI systems.

*****