A United Nations (UN) advisory group focused on artificial intelligence (AI) has urged governments to work together in establishing a “globally inclusive and distributed architecture” for regulating the technology.
Formed in October 2023 to evaluate international governance related to AI, the UN’s High-level Advisory Body on AI has released a report detailing its framework for mitigating AI-related risks and facilitating the equitable distribution of its transformative capabilities worldwide. The executive summary emphasized, “The need for global governance is undeniable. AI’s essential components, from critical minerals to training data, are sourced globally. The reach of general-purpose AI, utilized across borders, leads to diverse applications worldwide. The rapid development of AI is consolidating power and wealth on an international scale, with significant geopolitical and geo-economic repercussions.”
The report also pointed out the limitations in our understanding of AI’s complexities, stating that “no one fully grasps AI’s mechanisms to control its outputs or forecast its future developments. Decision-makers are often unaccountable for the creation, deployment, or use of systems they do not comprehend, which can lead to negative global consequences.”
While acknowledging that national and regional entities will play a key role in AI oversight, the Advisory Body argued that the inherently transboundary nature of AI necessitates a coordinated global effort. To this end, it has put forward several recommendations aimed at bridging the current gaps in AI governance. The Advisory Body is “calling upon all governments and stakeholders” to support the protection and development of human rights.
Key proposals include the establishment of an independent international scientific panel on AI, comprising a diverse range of multidisciplinary experts who would serve voluntarily, along with initiating a bi-annual intergovernmental and multi-stakeholder dialogue on AI governance at existing UN meetings.
Additionally, the report suggests forming an “AI standards panel” to maintain a register of definitions and standards for assessing AI systems, setting up a “global fund for AI” to help mitigate disparities, and creating a global AI data framework to promote robust local AI ecosystems.
Wendy Hall, a regius professor of computer science at the University of Southampton and the only British member of the Advisory Body, stated that current global measures for AI governance are inadequate. “We are witnessing AI advancements affecting health care, energy management, food production, and education,” she said. “Artificial intelligence must advance for the betterment of humanity, but without governance, it risks developing harms for society. At the very least, we hope this report sparks meaningful discussions regarding global governance. We urgently need a global framework that addresses the challenges posed by AI and ensures it benefits all nations, not just those at the forefront of its development,” Hall added.
In April 2024, a UN draft resolution advocating for the use of AI in sustainable development and human rights protection garnered support from over 120 member states. The resolution aimed to “bridge the artificial intelligence and digital divides within and between countries” by urging collaboration among member states, the private sector, civil society, research entities, and the media to develop regulations and governance frameworks for safe and trustworthy AI systems. However, this resolution was non-binding.
During the January 2024 World Economic Forum in Davos, UN Secretary-General António Guterres criticized technology companies and governments for prioritizing their narrow interests in AI “with a blatant disregard” for human rights and social implications, likening unregulated AI development to the escalating climate crisis. He called on global business and political leaders to rein in the “runaway” pace of AI development and advocated for reform in the post-World War II international system to better tackle global issues.
Highlighting the significant power imbalances in the current international framework as a barrier to progress in both AI governance and climate efforts, Guterres noted that many nations involved are remnants of colonial rule and therefore hold “minimal weight” in contemporary discussions.
The AI Advisory Body’s report acknowledged the abundance of frameworks, principles, and guidelines focused on AI governance that have emerged over recent years. However, it pointed out that none are globally representative, leading to issues with representation, coordination, and implementation. With many regions excluded from international AI governance discussions, it noted that only seven countries—Canada, France, Germany, Italy, Japan, the UK, and the US—are participants in non-UN AI governance initiatives, while 118 countries, mainly from the Global South, are not involved at all.
“Equity demands that diverse voices play significant roles in decisions regarding technology governance,” the report asserted. “The concentration of decision-making within the AI sector is indefensible; meanwhile, many communities have historically been excluded from discussions that directly impact them.”
To foster more equitable international collaboration on AI governance, the Advisory Body proposed creating a streamlined and agile structure: an AI office within the UN Secretariat, positioned close to the Secretary-General. This office would serve as a cohesive force to efficiently unite and sustain the various initiatives recommended in the report.
