“The United Nations is uniquely positioned to support the establishment of global principles and practice arrangements for AI training and data governance and use.”
– 2024 Report on Governing AI for Humanity
The UN report Governing AI for Humanity, published in September 2024, presents readers with numerous AI-related concerns, including the problem of governance, risks, automation, trust, and accountability. Though several AI patchwork initiatives do exist at present, the report identified numerous unresolved problems, including issues regarding representation, coordination, and implementation, which it sought to highlight and resolve.
An Independent Body
The report presents readers with a solution in the form of a proposal modelled on existing UN bodies such as the Intergovernmental Panel on Climate Change – an independent UN AI body that could serve as a “dedicated focal point” and “clearing house” for AI standards.
A Proposed Solution
In support of this solution, the report presents seven recommendations – essentially proposed responsibilities and tasks that a UN AI body could tackle, solutions it could present (and how it would present them), and the outcomes it would aim to achieve.
The seven recommendations presented in the report are:
- An International Scientific Panel on AI
- Policy Dialogue on AI Governance
- AI Standards Exchange
- Capacity Development Network
- Global Fund for AI
- Global AI Data Framework
- AI Office Within the Secretariat
Recommendations For Purpose and Direction
1. An International Scientific Panel on AI
A voluntary independent scientific panel of multidisciplinary experts who would prepare reports on an ad hoc, quarterly, and annual basis on the topic of AI and its impact on sustainable development goals (SDGs), risks, opportunities, trends, and more.
2. Policy Dialogue on AI Governance
Twice-yearly discussions on AI governance, best practices, AI-related incidents, and the latest findings and learnings from public and private sector specialists.
3. AI Standards Exchange
Bring representatives from standards development organisations, tech firms, civil society, and scientists to create and maintain a register of definitions and standards to measure and evaluate AI systems, the process for creating them, and identify where new standards are needed.
4. Capacity Development Network
Foster networking and resource- and education-sharing between countries to promote opportunities, including technology infrastructures, training and educational materials and courses, and fellowship opportunities with academic institutions and technology firms.
5. Global Fund for AI
Create and manage a mixed private/public fund to minimise the AI divide between countries and create opportunities to meet SDG targets, share resources for computing, modelling, testing, and benchmarking, and identify best practices and global governance and safety solutions.
6. Global AI Data Framework
Establish best practices and definitions around AI training data, including transparency around data provenance – ensuring all data is trustworthy. It also advocates for rules-based data exchanges and standardised contracts and protocols for accessing data.
7. AI Office Within Secretariat
The coordination body would report to the UN Secretary-General and would be staffed by UN personnel from specialised agencies who would engage with private, civil, and academic stakeholders, partner with leading external organisations, and foster common ground, mutual understanding, and shared benefits for all parties involved.
Seeking Common Ground
As laid out in the report, this proposed UN body for AI would seek to know and identify where and when it could provide coordination assistance or assume a leadership responsibility and when it should step aside and let others take the wheel.
An immediate and obvious benefit that comes from the creation of such a body within the UN is the existing set of institutional arrangements, pre-negotiated funding, and the well-oiled administrative processes available and at the disposal of the UN.
What Kind of Future
In the view of the report’s authors, technology such as this cannot be left to market forces and private, profit-driven organisations to govern. A new social contract, in the report’s view, is needed, and to that end the authors advance the idea of a “global buy-in” around governance, to ensure fairness, an even distribution of opportunities, and a minimisation of risk to the most disadvantaged and to future generations.
