Uncharted Skies of Liability: AI’s Role in Reshaping Aviation Accident Jurisprudence for Autonomous Aircraft Era

Introduction

Tracing the advent of autonomous aircraft: The AI revolution

“Technical advancement” since the very advent of structuring the legal framework to regulate and establish liability within the realm of “aviation accidents” has played the role of both progress catalyst and complexity harbinger. The 21st century AI revolution and the evolution of autonomous aircraft have introduced a dynamic technological sophistication for aviation operations. This unprecedented era not only holds promise for enhanced efficiency, improved safety, and innovation, but it also unleashes intricate legal challenges in the domain of “dispute resolution” and “liability” in accidents involving AI-operated aircraft.

Adding life to AI: The conscious automation

The nebulous area of artificial intelligence is constantly evolving, raising new questions of a “shared responsibility” among the various stakeholders involved. However, adopting a coherent theory of externally applied consciousness, a democratised attribute that goes beyond humans that makes machines “humane”, is an essential component of this automation, without which AI is nothing but age-old statistics.

This essay is an attempt to provide a sophisticated explanation of the intricate interplay between aviation accident laws and international conventions. It adds to the extending dialogue that identifies the need to adjust the legal framework compatible with the 21st-century technical frontiers while acknowledging the complex ethical, legal and policy issues underlying accidents involving autonomous aircraft.

The traditional liability and conventional challenges

International conventions, laws of torts and contractual agreements shape the liability in conventional aviation accidents. Primarily negligence is the guiding principle determining liability where harm results from failing to ensure reasonable care. In aviation operations, “human pilots” assess the dispute and liability appropriate based on several factors such as mechanical failure, human error, or negligence in air traffic control.

Distinguishing the liability framework in human-piloted and autonomous aircraft

AI-controlled aircraft have shifted the central focus from human error to mechanical autonomy and the intricacies of decision-making. A multifaceted liability is involved in AI algorithms and real-time data analytics owing to the decentralised nature of decision-making. The traditional concept of negligence does not apply to AI systems as, unlike human pilots, they lack volition and consciousness, therefore, calling for an elaborate evaluation of factors like adequacy of safe-fail mechanisms, system redundancies, quality of sensors and the AI programming. Since programming, designing, and testing autonomous aircraft are substantial to the operational liability, the role of manufacturers becomes pivotal.

The 21st-century compatible reforms

Integrating AI in the evolution of aviation technology requires fundamentally restructuring the liability framework in aircraft disputes and accidents. This paradigm shift entails a progressive regime from a singular outlook to human negligence to a comprehensive evaluation of the ecosystem involved in AI aviation operations. The reforms include new standards for care and reconsidering allocating a shared responsibility with an effective mitigation and risk-management mechanism. The international conventions designed to address human-piloted aircraft need revisions and amendments to address the complexities of autonomous air vehicles.

The existing legal jurisprudence

Critical study of the international treaties

The aviation industry has seen tremendous growth and so did the parallel laws to ensure a uniform regime for liability which protected the rights of the passengers and made the airline responsible and obligated for the passenger’s safety. The laws also provided the airlines with certain rights and privileges that would aid the development of the aviation industry. Thus, the legal framework led to a coexisting trust between the service provider and the consumer through a contractual relationship.

The first multilateral international treaty in the field of civil aviation law was “the 1919, Paris Convention¹”, this convention led to the creation of the International Commission for Air Navigation and is the fundamental structure for future treaties.

The Warsaw Convention, 1929² established the Comité International Technique d’Experts Juridiques Aériens and is ratified by 152 nations. The 1929 Convention has provided a uniform procedure for documentation and claims in case of damages. In case of accidents the airline is liable however, only limited liability is imposed as the Convention is inclined towards the interest of the airlines that could have ensured the expansion of the industry. In the 21st century where the wings of the aviation industry are spread wide across following the standards of the Warsaw Convention seems biased and unfair towards those who bear the harm. To rectify the liability caps and limits of the Warsaw Convention, in 1955 the Hague Protocol was passed. However, the Protocol could not result in significant changes as it also favoured the carriers similar to the Warsaw Convention and it was applicable only to the signatories of the conventions, furthermore, the United States is yet to ratify it.

The Montreal system, which replaced the Warsaw system under the international contractual liability regime, was modernised and amended to suit the expanding aviation industry better. Under the Montreal system, passengers and airlines have rights and protection. At the same time, airlines are subject to certain obligations as providers of air transportation services to passengers and avail several defences. The 1999 Convention establishes two distinct categories of liability: restricted or strict liability and unlimited liability. Under the strict liability regime, in the event of a passenger’s death or injury, the air carrier must pay a maximum of 100,000 special drawing rights, equivalent to 134,482 USD. Furthermore, under the Convention, the carrier must pay compensation in case of harm caused by delay, which is 4150 SDR.

Analysing the Indianised laws

Given the expanding aviation industry and India’s status as one of the world’s largest aviation markets, the Carriage by Air Act, 1972³ in India regulates the contractual relationship between airlines and passengers. The Convention for the Unification of Certain Rules Relating to International Carriage by Air⁴, the Hague Protocol of 1955⁵, and the Montreal Convention of 1999 were all given legal force by the passage of the Act. The 1972 Act implemented the aforementioned Conventions’ and protocol’s norms in their original, relevant forms. According to the Convention’s provisions, one can file a case. The compensation is also calculated as per the terms of the Convention, although there is no consistency in the computation of the compensation due.

Indian judicial perception

The Mangalore Air Crash⁶ involving an international airline crash where 158 were killed and 10 injured people including the crew was caused due to a pilot error. The Kerela High Court while hearing the writ petition to increase the compensation held that “Even though the actual damages payable under the law may be small in this scenario, the carrier should offer a reasonable minimum out of goodwill to avoid unnecessary lawsuits.” In Kandimallan Bharathi Devi v. General Insurance Corpn. of India⁷ it was established that although various restrictions and limitations are imposed on air companies under specific provisions of the Carriage by Air Act, 1972 such as the Third Schedule, in their defence against compensation claims for passenger harm or death, the general tort law that serves as the foundation for calculating damages is still applicable.

Aviation accidents and the doctrine of res ipsa locquitur

The California Supreme Court stated in Smith v. O’Donnell⁸ that when it clear that the occurrence was not an ordinary incident without negligence by those involved in the function and the event due to which injury occurred was in charge of the party who is sought to be charged, and there was no negligence which the defendant can prove, then a presumption of negligence is raised which the defendant must overcome by proof that there was, in fact, no negligence.

The carrier’s liability

The US Court of Appeals recently resolved the debate over how to interpret Article 17⁹ in Prescod v. AMR, Inc.¹⁰ In this instance, airline staff members took a senior citizen’s suitcase. Other drugs, as well as life-supporting breathing apparatuses, were discovered in the bag. The court found that the senior citizen’s suitcase being taken away based on a single suspect constituted an “unusual or unexpected event or happening external” to the passenger. Therefore, for Article 17 of the Warsaw Convention, this act of the carrier was declared an “accident”.

Quantifying emotional damages

Legal precedent in the United Kingdom has grown unambiguous regarding emotional distress. Lord Steyn ruled in Morris v. KLM Royal Dutch Airlines¹¹ that any suffering brought on by physical harm sustained in the flight disaster is compensable. In addition, the Court made clear that “where a relevant accident results in mental injury or disease that results in undesirable physical symptoms”.

Vicarious liability and aircraft disputes

The legality of an aircraft accident that results in the death or injury of a passenger, as well as the liability of the aircraft owner, lessor, or secured party, is still being disputed. Depending on the jurisdiction, a secured party, lessor, or owner of an aircraft are liable even if they did not have absolute ownership or control of the plane at the time of the event or accident.

Towards the shaping of new legal standards

On the surface, it appears natural to equate an AI creature to an aeroplane flying on autopilot while a pilot remains in the pilot’s seat. In Brouse v. United States¹², decided in 1949, the Ohio District Court held a pilot utilising autopilot mode liable for a disaster involving two aeroplanes. The Court stated that those in charge of robotically controlled aeroplanes must maintain a “proper and constant lookout”, especially when the plane is flying through recognised and established airways. However, this established principle when applied to the AI aviation industry proves to be inadequate due to primarily three considerations:

(a) Firstly, we cannot be certain that the future autonomous aircraft models operated by AI will even have human involvement.

(b) Secondly, liabilities of army airplane pilots and commercial pilots differ.

(c) Finally, the airplane operated in the sky is distinct from the “robotic” or “AI” controlled devices on the sea or ground.

Time constraints and predictable payouts are valuable tools in the hands of the aviation industry; nevertheless, they are inappropriate in the AI setting. Assuming AI entities are analogous to autonomous vehicles, time-limits are inapplicable to AI entities since imposing the same stringent time constraints scheme to use autonomous cars on the ground or autonomous vessels at sea is unjust and unrealistic.¹³

When we leave the aerial area, the certification requirements are radically different and less stringent, and the physical infrastructures required for the efficient participation of many vehicles on the ground are vastly different from those needed for aeroplanes in the air. Strict liability with a built-in roof restriction may be applicable to autonomous entities on the ground under the payment approach.

It will be difficult to assert that the AI entity’s operator should be held accountable instead of the AI device’s manufacturer unless the operator has done something to deflect the harm suffered from the AI entity itself to the operator’s behaviour. In Tempe, Arizona, in March 2018, an Uber autonomous car crashed, killing pedestrian Elena Herzberg. At the time, a human driver was in the driver’s seat.¹⁴ This accident provides a modern illustration of this idea. After an accident, it was held through an investigation that the automobile had seen the pedestrian but had not actively tried to stop or warn the driver despite being aware of them. These factors suggest that it is more challenging to hold the machine operator liable than the developer or maker because the former’s significant rule is monitoring the AI entity’s behaviour instead of dictating or directing it.

Traditionally, the aviation industry’s safety standards, certification procedures, and operational guidelines have been established by regulatory agencies. In the age of AI, their responsibilities also include:

(i) Evaluating the dependability and safety of AI algorithms.

(ii) Setting performance standards.

(iii) Carefully examining the incorporation of AI systems into aviation.

Regulatory organisations must promote a thorough grasp of AI technology and its consequences for aviation safety to settle disputes and obligations.

Ethical issues are critical in adapting legal norms to the AI era. AI algorithms’ decision-making processes create ethical concerns because they lack consciousness, deliberateness, and moral autonomy. If AI systems render life-critical judgements, regulatory agencies must contend with issues of transparency, accountability, and the possibility of algorithmic bias.

A comprehensive strategy encompassing technological innovation, ethical issues, and nuanced knowledge of legal requirements is essential as aviation regulatory agencies adjust to the AI era. A seamless integration of these variables is necessary due to the interaction between developing AI technologies and the complexity of aviation law. Regulatory bodies can guarantee that the regulatory structures concerning AI-controlled aircraft strike an adequate balance between security, innovation, and ethical accountability in the dynamic environment of the 21st century by actively engaging with technological breakthroughs and stimulating open ethical standards.

The sky of future: Concluding remarks and suggestions

The deadly Boeing 737 Max Crashes¹⁵ caused due to the alleged changes to the AI software Maneuvering Characteristics Augmentation System (MCAS) that were aggressive and risky provide an insight into the long-term potential implications of the role played by the artificial intelligence in the aviation accident jurisprudence and initiates the need of collaborative efforts of both legal and technical experts.¹⁶ Conflicts and liabilities in the aviation industry must occur in an era where AI is already part of a changing society. AI can also be easily tricked or manipulated and is also quite fragile. What is referred to as AI-explainability is the ability to understand why software code makes certain judgements and to build methods to make AI more resilient in this situation. Without the integration of ethics in AI the development of AI, intelligence and machine learning cannot be sustained in a world where machines and humans have to coexist and share a codependent relationship.

Another incident where Shady¹⁷, $1.5 million military drone due to a bug without human supervision crossed three States and mountains throws light on both the advantages and risks of embracing AI in aviation be it civil or military operations as it can be both terrifying and rewards at the same time.

A thorough rethinking of the culpability paradigm is required due to the advancement of aviation technology, especially the integration of AI. The complexity of accidents utilising autonomous systems is difficult for the traditional negligence-based approach to handle appropriately. To ensure accountability, safety, and innovation within the aviation industry, a liability system that looks to the future must accept the unique qualities of AI-controlled aircraft. To meet the opportunities and problems posed by the dynamic environment of AI aviation in the twenty-first century, an upheaval in liability architecture and regulatory requirements is needed.