The UN's new draft resolution on AI encourages Member States to implement national regulatory and governance approaches for a global consensus on safe, secure and trustworthy AI systems.

Laws/Regulations directly regulating AI (the “AI Regulations”)

The United Nations (UN) adopted a draft resolution on AI, entitled "Seizing the opportunities of safe, secure and trustworthy artificial intelligence systems for sustainable development" (the "Resolution") on
March 21, 2024. The Resolution was backed by more than 120 Member States and aims to encourage countries to safeguard human rights, protect personal data and monitor AI for risks on a non-legally binding basis.¹ It was designed to complement other work done by the UN, including the International Telecommunication Union (ITU), the UN Educational, Scientific and Cultural Organization (UNESCO) and the Human Rights Council.

Status of the AI Regulations

The UN does not have the ability to pass laws or regulations regarding AI or its implementation. However, the UN Charter gives the General Assembly the power to initiate studies and make recommendations to promote the development and codification of international law. In the future, the General Assembly may vote on other AI resolutions, which are expressions of the Member States' views and are not legally binding.

Other laws affecting AI

The Resolution aims to reach a global consensus on safe, secure and trustworthy AI systems, and encourages its Member States to implement and develop their distinct national regulatory and governance approaches and frameworks for the full realization of the 2030 Agenda for Sustainable Development.²

In addition, there are various existing frameworks that do not directly seek to regulate AI but may affect the development or use of AI in the UN. For example:

• International human rights law continues to apply (e.g., the Charter of the United Nations and the Universal Declaration of Human Rights etc.)
• Private sector activities of all AI actors should comply with domestic laws and international frameworks, such as the United Nations Guiding Principles on Business and Human Rights: Implementing the UN "Protect, Respect and Remedy" Framework and the OECD Guidelines for Multinational Enterprises

Definition of “AI”

The Resolution does not establish an independent definition of "AI." Instead, it sets out "the safe, secure and trustworthy artificial intelligence systems" that refer to "artificial intelligence systems in the non-military domain, whose life cycle includes the stages: pre-design, design, development, evaluation, testing, deployment, use, sale, procurement, operation and decommissioning, and are such that they are human-centric, reliable, explainable, ethical, inclusive, in full respect, promotion and protection of human rights and international law, privacy preserving, sustainable development oriented, and responsible —and have the potential to accelerate and enable progress towards the achievement of all 17 Sustainable Development Goals and sustainable development in its three dimensions —economic, social and environmental— in a balanced and integrated manner; promote digital transformation; promote peace; overcome digital divides between and within countries; and promote and protect the enjoyment of human rights and fundamental freedoms for all, while keeping the human person at the centre".³

Territorial scope

The Resolution does not have a specific territorial scope. Its primary goal is to ensure "safe, secure and trustworthy AI systems" on a global level. It encourages all 193 Member States and multi-stakeholders from all regions and countries (including from the private sector, international and regional organizations, civil society, the media, academia and research institutions and technical communities and individuals) to develop and support regulatory and governance frameworks.⁴

Sectoral scope

The Resolution is not sector-specific. It is aimed at all sectors and actors on a global level.

Compliance roles

The Resolution encourages all Member States and stakeholders to develop and support the realization of "safe, secure and trustworthy AI systems", but does not designate explicit compliance roles.

Core issues that the AI Regulations seek to address

The Resolution seeks to:

• Bridge the AI and other digital divides between and within countries
• Promote safe, secure and trustworthy artificial intelligence systems
• Encourage Member States and all stakeholders within their roles and responsibilities, support and develop regulatory and governance frameworks for safe, secure and trustworthy AI systems that create an enabling ecosystem at all levels
• Call on Member States and other stakeholders to take action and assist developing countries toward inclusive and access to the benefits of digital transformation
• Emphasize that the human rights and fundamental freedoms must be respected, protected and promoted throughout the life cycle of artificial intelligence systems; call on all Member States and, where applicable, other stakeholders to refrain from or cease the use of artificial intelligence systems that are impossible to operate in compliance with international human rights law
• Emphasize the fair, inclusive, responsible and effective data governance
• Encourage the private sector to adhere to applicable international and domestic laws and act in line with the United Nations Guiding Principles on Business and Human Right
• Call on specialized agencies, funds, programs or other entities to assess and enhance their response to leverage the opportunities and address challenges posed by AI systems⁵

Risk categorization

AI is not explicitly categorized according to risk in the Resolution.

Key compliance requirements

The Resolution does not impose compliance requirements but calls on Member States and other stakeholders to take action and assist developing countries toward inclusivity and access to the benefits of digital transformation, including by:

• Expanding participation of all countries in harnessing the benefits of digital transformation (increasing digital literacy, closing the digital divide)
• Enhancing digital infrastructure connectivity and access to technological innovations through stronger partnerships to help developing countries effectively participate throughout the life cycle of artificial intelligence systems, and accelerate the inclusive and positive contribution of artificial intelligence systems to society
• Enhancing the ability of developing countries to address major structural impediments to accessing the benefits of new and emerging technologies and AI innovation
• Aiming to increase funding for Sustainable Development Goals –related research and innovation (for digital technologies)
• Enabling international innovation-based environments for developing countries to develop technical expertise and capabilities
• Urgently mobilizing means of implementation such as technology transfer on mutually agreed terms
• Promoting access and design, development and use of safe, secure and trustworthy AI systems⁶

The Resolution also encourages Member States and other stakeholders to (for the benefit of all) address the world's greatest challenges and achieve sustainable development in economic, social and environmental dimensions (with specific consideration of developing countries) by:

• Promoting the development and implementation of domestic regulatory and governance approaches and frameworks
• Encouraging effective measures that promote innovation (to identify, classify, evaluate and mitigate risks)
• Encouraging incorporation of feedback mechanisms (for discovery and reporting of misuse of AI systems)
• Raising public awareness and understanding of the core functions of AI systems
• Fostering the development, implementation and disclosure of risk monitoring and management mechanisms
• Strengthening investments in developing and implementing effective safeguards (like risk and impact assessments, security systems)
• Encouraging the development and deployment of effective and internationally interoperable technical tools and standards or practices to identify information manipulation
• Facilitating the development and implementation of effective, international interoperable frameworks and standards for training and testing AI systems
• Encouraging the implementation of appropriate safeguards for intellectual property rights
• Safeguarding privacy and protection of personal data
• Promoting transparency, predictability and reliability throughout the life cycle of AI systems that make or support decisions impacting end-users
• Promoting AI systems that preserve and protect linguistic and cultural diversity
• Intensifying information sharing on mutually agreed terms among entities within roles across an AI system life cycle⁷

Regulators

As the Resolution has no enforcement powers on its Member States, there are no regulators under the Resolution, nor does the Resolution stipulate how the Member States should regulate AI systems in their own jurisdictions.

Enforcement powers and penalties

As the Resolution is not legally binding, it does not confer enforcement powers or give rise to any penalties for non-compliance.

_{1 See the UN News Article and the Resolution.
2 See the Resolution, pages 3 and 4.
3 See the Resolution, page 2.
4 See the Resolution, pages 3 and 4.
5 See the Resolution, pages 3 to 7.
6 See the Resolution, pages 4 and 5.
7 See the Resolution, pages 6 and 7.}

Jeffrey Shin (Trainee Solicitor, White & Case, London) contributed to this publication.