Implementation of Artificial Intelligence in International Aviation: A Critical Study

Artificial Intelligence (AI) will provide the aviation sector with $8 billion in operational efficiencies by 2022. The implementation of AI in Aviation could minimize human-error and other risks. This paper investigates the legal viability of the implementation of Artificial Intelligence in Commercial Aviation, it further includes the examination of the central structural treatise on the topic, the study of regional Bilateral Aviation Agreements, and Multilateral Agreements (Mas). To identify the lapses of the current laws and identify policy-spaces. The paper contains a review of the existing literature, analysis of key trends and developments of international law, and a corelation with the International Law harmonization efforts. Part I of the paper lays down the definition of AI and its scope in aviation. Part II of the paper investigates the International Legal Framework for Aviation, existing legal regimes and ‘Structural-Regimes’ of International Aviation Law, Bilateral Investment Treaties and MAs. Part III investigates the existing liability mechanisms and its viability for suitable implementation of AI into Aviation. Part IV presents Data-Protection and Privacy related concerns arising out of AI implementation. Part V supplements the paper with the most relevant case studies. Part VI presents the future trends and challenges into the implementation of AI in aviation. Part VII of the paper presents the conclusions and the policy recommendations. The study concludes, based on the analysis of the relevant aviation law, that the way forward is to implement AI related nomenclature and Data-Privacy related specific concerns as ‘minimum standards’ in the model BITs and MAs, as a part of the harmonization efforts so that the liability regimes and multiple other forms of dilemmas could be effectively resolved. There should be a centralized legal framework consisting of the major aviation stakeholders a regulatory authority. Every automation effort should have a human supervisor to minimize overall risks and dilute the AI liability dilemma.