2025-06-29-联合国贸易发展委员-联合国贸易发展委员会-全球合作促进包容性和公平的人工智能（英）_9页_644kb

Summary of Policy Brief: Global Collaboration for Inclusive and Equitable Artificial Intelligence

Core Content

This policy brief highlights the need for global collaboration to ensure inclusive and equitable artificial intelligence (AI) development. It emphasizes the uneven distribution of power and benefits in AI, particularly the dominance of multinational technology giants and the limited representation of developing countries in international governance. The brief proposes a multi-stakeholder approach and key initiatives to address these challenges.

Main Points

1. Dominance of Multinational Technology Giants

• The private sector, especially multinational tech companies, drives AI innovation and development.
• These companies often operate in an oligopolistic environment, with significant market power and resources.
• Examples include Alphabet's acquisition of DeepMind, Microsoft's partnership with OpenAI, and Nuance Communications' acquisition by Microsoft.
• These firms control large datasets, cutting-edge research, and frontier machine learning models, with public institutions and academia lagging behind in both quantity and quality of AI outputs.
• The market dominance of these companies can lead to suboptimal development paths, replacing human labor, and neglecting ethical considerations such as bias, misinformation, and transparency.
• Antitrust actions are being taken in various jurisdictions, but developing countries often lack the institutional capacity to enforce regulations effectively.

2. Lack of Representation of Developing Countries in Global AI Governance

• The current AI governance framework is fragmented and led by developed nations, particularly Group of 7 (G7) members.
• 118 countries, mostly from the Global South, are not part of any major AI governance initiative.
• This imbalance undermines the legitimacy and effectiveness of global AI governance.
• The United Nations has taken steps to address this, including:
  • Adopting resolutions on safe and trustworthy AI systems for sustainable development.
  • Committing to international cooperation in capacity-building.
  • Establishing an Independent International Scientific Panel on AI.
  • Initiating a Global Dialogue on AI Governance.
  • Requesting the Commission on Science and Technology for Development to create a dedicated working group on data governance.

3. Need for a Public Disclosure Mechanism

• A public disclosure mechanism is proposed to enhance accountability and ensure tangible outcomes from global AI commitments.
• This mechanism could be based on the ESG reporting framework, with impact assessments throughout the AI life cycle.
• It should include transparent reporting on how AI systems function, data collection and management, and environmental and social impacts.
• The mechanism should evolve from voluntary to mandatory as standards mature.
• Balancing innovation with public safety and trust is crucial, avoiding both overregulation and underregulation.

4. Need for International Cooperation in Digital Infrastructure, Data and Skills

• The three key drivers of AI advancement are:
  • Digital infrastructure (computational power and cost-effective information transfer).
  • Data (volume, diversity, and growth).
  • Skills (ranging from basic literacy to advanced expertise).
• International collaboration is essential to ensure equitable access and shared benefits.
• It can help developing countries overcome challenges such as inadequate infrastructure, limited data access, and skills shortages.
• Collaboration can prevent fragmentation, duplication of efforts, and increased inequalities.

Key Areas for Collaboration

Shared Digital Public Infrastructures

• A global shared facility, akin to CERN, could provide equitable access to AI infrastructure.
• Public-private partnerships can accelerate the development of local innovation ecosystems.
• Tailored systems can offer essential resources and services for AI adoption.

Open Innovation

• Open data and open source models can democratize knowledge and resources.
• Harmonizing open-source efforts and adopting common standards can improve global knowledge sharing and access.
• The Manaus package from the G20 Research and Innovation Working Group includes an open innovation strategy and guidelines for inclusive AI development.

Capacity-Building Initiatives

• A global AI hub, similar to the United Nations Climate Technology Centre and Network, can support capacity-building in developing countries.
• Regional innovation hubs and expert networks can strengthen South-South cooperation and address common AI challenges.

Conclusion

To achieve inclusive and equitable AI development, international collaboration is essential. This includes:

• Engaging developing countries in AI governance.
• Establishing a public disclosure mechanism for accountability.
• Promoting shared digital infrastructure, open innovation, and capacity-building.

These measures can help align AI development with shared global goals, ensure ethical and transparent practices, and mitigate inequalities.

Contact Information

• Torbjörn Fredriksson
  Officer-In-Charge, Division on Technology and Logistics, UNCTAD
  Email: Torbjorn.Fredriksson@unctad.org
  Phone: +41 22 917 2143

• Angel Gonzalez Sanz
  Officer-In-Charge, Division on Technology and Logistics, UNCTAD
  Email: Angel.Gonzalez-Sanz@unctad.org
  Phone: +41 22 917 5508

• Press Office
  Email: unctadpress@unctad.org
  Phone: +41 22 917 5828
  Website: unctad.org