Regulating Artificial Intelligence under International Law: Emphasizing the Necessity of Global Treaty

Abstract: Currently AI has been incorporated in various industries like healthcare, security and finance. By 2030, it is expected to add upto 15.7 trillion dollars to global economy[1]. In the recent times, AI is being empowered using such advanced algorithms that it can make decisions on its own and use force on the enemy without human guidance[2]. This creates a risk of AI bypassing current regulatory frameworks, thus, being a persistent threat to life, privacy and intellectual property risk and liability. In the present scenario, there is the lack of a dedicated AI regulatory system. Global AI regulation remains a combination of how regulatory statutes of several states. The European Union AI Act is more focused on human rights while the frameworks of US and China are market oriented and state centric[3] respectively. Also, these frameworks are largely dependent on non-binding or quasi legal instruments like the Organisation of Economic Co-Operation and Development (OECD) AI principles and the Recommendation on the Ethics of AI as given by UNESCO[4]. Although, the ‘Council of Europe Framework Convention on Artificial Intelligence and Human Rights, Democracy, and the Rule of Law’, the first binding treaty agreed upon by US, EU and other States has been passed, it fails to answer the question regarding enforcement of the same[5]. Problems of algorithmic bias, privacy and security infringement, accountability, transparency and protection of fundamental rights are largely attributed to the regulatory vacuum. The aim of this article is to explore the extent to which a binding treaty is needed for the cause of addressing the accountability and responsibility gap created.

1. Existing International AI Frameworks : UNESCO Recommendations and OECD Frameworks

• UNESCO AI Ethics Recommendations –

This framework stands as one of the most normative frameworks till date. In 2021, it was unanimously adopted by 193 Member States. Largely based on principles governing human dignity, environmental sustainability, non discrimination and equality, it reflects human rights centric approach. To ensure accountability, it introduces monitoring obligations like periodic reporting and suggests the establishment of National AI Ethics Commissions[6].

• The OECD AI Principles

This is the first intergovernmental consensus on AI ethics and governance as adopted on 2019 and amended in 2024. It has helped in the formation of national strategies to regulate AI in several countries like Japan, US and Canada besides impacting and acting as a guideline for G20 AI principles. It has systems like AI policy Observation which is aimed to support the enforcement and aid peer learning[7].

• Limitations of Current Frameworks –

• These regulations can be explained as Quasi-Legal Systems, given their non binding nature. The issue that comes forward due to such a characteristic is the primarily the fact that they would not have the authority needed to guarantee the abidance to such safety standards as prescribed[8].
• Since compliance to them becomes optional and overly dependent on political or corporate will, the problem of non uniformity creeps in.
• In case of violation of international AI norms, there is no Global authority that would investigate, put sanctions on defaulting private or public sector authorities.
• While regional regulations like those of EU and China impose penalties, globally the lack of enforcement measures persist[9].
• A lot of countries, like India, still depend on traditional and procedural bureaucratic supervision methods. In such a scenario, the possibility of human intervention not being able to keep pace with the growth of AI cannot be ignored[10].
• Also, since a uniform global framework does not exist, there stays the possibility for different jurisdictions adopting frameworks that may differ in level of strictness. This opens the ground for global firms to shift to countries where AI regulatory laws do not require compliance with rigorous or strict adherence to prescribed global standards.

2. Challenges Under International Humanitarian Laws and Human Laws

• Conflict with International Humanitarian Laws

• Principles like distinction and proportionality were established considering that human decision and judgement formers would exercise the principles of responsibility and accountability while they evaluate incidental harm[11].
• While military sectors are planning the use of advanced algorithms that can make decisions as to the force used on enemy forces in warfare without human intervention[12], the question lingers as to whether AI would be able to make a distinction between civilians and security forces.

• Concerns regarding Human Rights

• AI systems can misuse the database they are based on. This database is chiefly formed of massive and indiscriminate personal data of millions of users.
• Also, since AI algorithms are formed by none other than human beings who inherently cannot be assumed to not possess individual biases based on caste, gender, skin colour etc and other factors, there remains a substantial doubt as to the algorithmic bias[13] it possess.
• The use of biological recognition tools by State without required supervision poses a significant threat to individual rights and freedom.
• The lack of transparency in case of use of AI creates barriers to accountability, which robs people of their right to clear or consistent pathways as to understand the point at which their human rights were violated.

3. The Need for a Dedicated AI Regulatory Framework

• Uniform Global Standards –

The current regional regulatory frameworks of EU, US and China need to be consolidated in a way that does not threaten the current landscape of international humanitarian laws and fundamental rights. This would ensure algorithmic inter–operability and facilitate compliance of multinational entities to a unified global framework in place of regional laws.

3.2 Ensuring Accountability and Enforcement –

The OECD principles and UNESCO AI Ethics Recommendations, collectively termed as ‘Soft Law’ owing to their quasi legal nature, fail to provide enforcement guidelines. A Global Treaty governing AI regulation would subsequently provide measurable penalties in instances of violation thus ensuring that accountability and responsibility vacuums cease to exist.

• Preventing an AI Arms Race –

A Global Treaty can be expected to ensure the moderation of the competition of global powers and enhance focus on technical safety. It could also help to specify the authority as to who can use and control critical AI Inputs like high performance chips. This could fairly prevent the proliferation of harm causing abilities of AI. Incorrect predictions and escalatory feedback loops can be effectively stopped through international and inter-governmental co-operation.

•  Strengthening Human Rights –

The presence of a global treaty will help to make AI such that it is used to supplement human capabilities rather than eliminating human judgement in the first place. Unethical AI practices like mass surveillance and social scoring which are discriminatory in nature and violate fundamental rights can be prohibited. Human dignity and the rule of law will stay unthreatened to quite an extent. Also, such a framework would protect the right to remedy as adequate transparency measures will be undertaken. This would further help in the enhancements of human abilities and welfare.

Conclusion –

The rapid advancement of Artificial Intelligence has exposed a fundamental inadequacy within the existing international legal order. While soft law instruments such as the OECD AI Principles and UNESCO Recommendations have laid an important normative foundation, their non-binding nature and lack of enforceability render them insufficient to address the complex and transnational challenges posed by AI systems. The fragmented regulatory approaches adopted by different jurisdictions further exacerbate this issue, creating inconsistencies, regulatory loopholes, and opportunities for misuse.

As AI systems increasingly operate with autonomy and cross-border impact, the absence of a unified legal framework risks undermining core principles of international law, including accountability, state responsibility, and the protection of fundamental human rights. The challenges posed to international humanitarian law, particularly in the context of autonomous weapons, and the growing concerns surrounding algorithmic bias, privacy violations, and lack of transparency, highlight the urgency of regulatory intervention at a global level.

In this context, the formulation of a binding international treaty emerges not merely as a regulatory option, but as a necessity. Such a treaty would facilitate the establishment of uniform global standards, ensure enforceability through defined accountability mechanisms, and prevent the unchecked proliferation of harmful AI capabilities. It would also play a crucial role in safeguarding human rights by embedding principles of transparency, fairness, and human oversight within AI governance frameworks.

Ultimately, the objective of international AI regulation must not be to hinder technological innovation, but to ensure that such innovation evolves within the boundaries of law, ethics, and human dignity. A global treaty, therefore, represents a critical step towards reconciling technological progress with legal responsibility, ensuring that Artificial Intelligence remains a tool for human advancement rather than a source of uncontrolled risk.

In addition, the development of a global AI treaty would foster greater international cooperation, ensuring that technological advancements are not monopolized by a few powerful states but are governed through shared responsibility. Such a framework would also encourage capacity building in developing nations, enabling them to participate meaningfully in AI governance and innovation. By promoting transparency, trust, and collaboration among states, a binding treaty can reduce geopolitical tensions and prevent regulatory fragmentation. Ultimately, a cohesive international approach will ensure that the benefits of AI are equitably distributed while minimizing its risks, reinforcing a balanced relationship between technological progress and global legal order.

Reference(S):

[1] African Union, Continental Artificial Intelligence Strategy, Notified on July 2024, https://au.int/sites/default/files/documents/44004-doc-EN-_Continental_AI_Strategy_July_2024.pdf

[2] ERICA H. MA, AUTONOMOUS WEAPONS SYSTEMS UNDER INTERNATIONAL LAW, 95 NYULJ 1435, 1437 2020

[3] Legalithm Team, AI Regulation Compared : EU, US, UK, China (2026), The Global Governance Deficit and the “Regulatory Mosaic” (April 7, 2026), https://www.legalithm.com/en/blog/ai-regulation-comparison-eu-us-uk-china-global

[4] Hiroki Habuka & David U. Socol de la Osa, Shaping Global Ai Governance : Enhancements and Next Steps for the G7 AI Process (24 May 2024), https://www.csis.org/analysis/shaping-global-ai-governance-enhancements-and-next-steps-g7-hiroshima-ai-process

[5] Jose-Miguel Bello Villarino et al., A first step on the long road to Global AI Regulation, https://www.lowyinstitute.org/the-interpreter/first-step-long-road-global-ai-regulation

[6] Sijuola Atanda-Lawal & Victor Chinonye Onuchukwu, The Global Governance Of Ai: Between Soft Law And Binding Obligations In International Law, 13 IJECM 159, 165-166, (2025)

[7] Sijuola Atanda-Lawal & Victor Chinonye Onuchukwu, The Global Governance Of Ai: Between Soft Law And Binding Obligations In International Law, 13 IJECM 159, 165-166, (2025)

[8] Hiroki Habuka & David U. Socol de la Osa, Shaping Global Ai Governance : Enhancements and Next Steps for the G7 AI Process (24 May 2024), https://www.csis.org/analysis/shaping-global-ai-governance-enhancements-and-next-steps-g7-hiroshima-ai-process

[9] Jose-Miguel Bello Villarino et al., A first step on the long road to Global AI Regulation, https://www.lowyinstitute.org/the-interpreter/first-step-long-road-global-ai-regulation

[10] Narendra Reddy, Towards Harmonized International Ai Governance: The United Nations In Ethical And Legal Systems, 5 IJIRL 1563, 1570 (2017)

[11] Narendra Reddy, Towards Harmonized International Ai Governance: The United Nations In Ethical And Legal Systems, 5 IJIRL 1563, 1570 (2017)

[12] ERICA H. MA, AUTONOMOUS WEAPONS SYSTEMS UNDER INTERNATIONAL LAW, 95 NYULJ 1435, 1437 2020

[13] Elif Kiesow Cortez, AI and Privacy Risks: The EU AI Act vs. US NIST AI Risk Management Framework, https://law.stanford.edu/wp-content/uploads/2024/11/TTLF-WP-125-Kiesow-Cortez.pdf