Abstract

This article examines the potential for international responsibility to arise for States due to their reliance on artificial intelligence (AI) systems in environmental forecasting within fragile Arctic ecosystems. This is particularly considering the limitations of existing legal frameworks, including the United Nations Convention on the Law of the Sea (UNCLOS) and the Convention on Biological Diversity (CBD), to address this technological development. The article assesses the possibility of attributing to the State decisions made based on AI outputs in consideration of the attribution rules outlined in the Articles on Responsibility of States for Internationally Wrongful Acts (ARSIWA) and relevant international judicial precedents. Through this analysis, the article distinguishes between attributing such decisions or omissions to the State and determining whether they breach international obligations. The findings indicate that relying solely on AI‐generated outputs without human evaluation or independent technical review may constitute a breach of the duty of due diligence and trigger international responsibility in the event of environmental harm. The article recommends the development of a binding international protocol to govern the use of AI in ecologically sensitive areas, ensuring transparency, periodic auditing and the protection of Indigenous rights.

Abstract

Scholars have identified challenges to protecting individuals from discrimination in contexts where organisations deploy artificial intelligence decision-making processes. While scholarship on ‘digital discrimination’ is growing, scholars have paid less attention to the impact of the use of artificial intelligence decision-making processes on persons with disabilities. This article posits that while the use of artificial intelligence technology can be beneficial for some purposes, its deployment can also construct a disability. The article demonstrates that the Convention on the Rights of Persons with Disabilities can be interpreted in a manner that confers a wide variety of human rights on persons with disabilities in the context when entities deploy artificial intelligence decision-making processes. The article proposes a test for digital discrimination based on disability and shows how it can be incorporated into the treaty through legal interpretation. Thereafter, it moves to developing an analogous general test for digital discrimination under international human rights law, applicable beyond a catalogue of protected characteristics.

Abstract

The Government's use of algorithmic-based decision-making is rapidly expanding across policy areas, including immigration, social security, regulation, security and policing. This book provides the first comprehensive analysis of how public law applies to the use of artificial intelligence and automation in the public sector in England and Wales. Starting with an accessible account of the nature of AI and automated systems being increasingly deployed in the public sector, the book covers the various legal regimes which regulate their use. It considers how the principles of judicial review might be deployed to challenge automated decision-making by public authorities. It also explains how equality law, human rights law, procurement law, data protection law and private law apply to government use of AI and automation. This book is a vital guide for practitioners in both private practice and government, and for anyone navigating this quickly changing, complex and uncertain environment.

Abstract

Abstract International law prohibits states from intervening in the internal and external affairs of other states, but only if the method of intervention is coercive. This Article argues that coercion can be understood in two different ways or models. First, as coercion-as-extortion, a demand coupled with a threat of harm or the infliction of harm, done to extract some kind of concession from the victim state—in other words, an act targeting the victim state's will or decision-making calculus. Second, as coercion-as-control, an action materially depriving the victim state of its ability to control its sovereign choices. This may be done even through acts like cyber operations that the victim state is entirely unaware of. The Article argues that many of the difficulties surrounding the notion of coercion arise as a consequence of failing to distinguish between these two different models.