AI governance requires a global approach that treats all countries as equal partners
In October 2024, the Prime Minister of India emphasized the need for a global framework for ethical Artificial Intelligence (“AI”). Various AI experts worldwide have made similar calls for international cooperation on AI governance. Several developments have taken place in this direction such as global AI summits, the signing of international declarations, and the setting up of AI safety institutes by various countries. This article highlights the importance and need for global governance of AI and related challenges.
Image generated by Microsoft Designer
Need for global governance
The development and use of AI technologies are not limited to one country. From the critical minerals needed to power AI technologies to the training data used in their development, these resources are sourced globally. Furthermore, the impact of AI on society transcends national borders. Recently, the Governing AI for Humanity report by the UN (“UN Report”), noted that the development and use of AI cannot be left to the markets. While national governments and regional organizations are important, the very nature of AI (transboundary in structure and application) requires a global approach to its governance. The Report also highlighted that a global governance approach for AI can be a key enabler of AI innovation for Sustainable Development Goals (“SDGs”) globally. Furthermore, several standards bodies such as the International Organization for Standardization and the International Telecommunication Union are coming up with standards governing AI. However, there is no common language among these bodies and many terms like fairness, safety, and transparency do not have agreed definitions. For instance, fairness in AI refers to the absence of bias or discrimination in AI systems. The grounds for discrimination vary by context. In the USA, factors like class and race play a central role. Meanwhile, in diverse countries like India, discrimination takes on different forms—shaped by religion, caste, language, ethnicity, and other social identities. Therefore, there is a need to ensure that there is global coordination with various stakeholders and cultures in creating such definitions and standards that enhance their relevance and effectiveness. This will ensure that core principles of safe, responsible, and ethical AI consider the unique needs and requirements of diverse demographics of the world.
Making global governance inclusive
One of the key considerations when thinking about global AI governance is ensuring that it is truly inclusive with equal participation from all countries. The UN Report studied seven prominent, non-UN AI initiatives. It found that seven countries were parties to all sampled initiatives, whereas 118 countries were parties to none and these were primarily from the global south. Such exclusion may lead to a fragmented global AI policy landscape where the interests and concerns of such countries are not incorporated in AI standards and governance frameworks. Therefore, it is imperative to ensure that we adopt an inclusive approach to AI governance that takes along all parts of the world including the Global South as equal partners. A similar call was made by several AI researchers and policy experts through the Manhattan Declaration. The declaration noted that the impacts of AI are global and the approach to its development and governance must be fair and inclusive. It emphasized the importance of including diverse perspectives from researchers worldwide regardless of their backgrounds, geographic locations, or institutional affiliations.
"AI's impact transcends national borders, affecting countries in diverse and unique ways. Global coordination on AI is crucial, but true inclusivity demands that the voices of the Global South are well represented. In a nation as diverse as India, with its multitude of languages, cuisines, and cultural heritage, it is vital that our distinct needs are incorporated into global AI regulations and standards. This approach will help ensure that AI supports equitable development and respects the diversity of our global community." - Prof. Balaraman Ravindran, Head of Centre for Responsible AI (CeRAI), IIT Madras
Operationalizing global coordination
While there is consensus among countries and AI experts for global coordination on AI, there is a need to operationalize this consensus to develop a shared understanding of AI safety, standards, and governance frameworks. The UN Report recommended setting up a framework for global governance for AI at the UN. It includes setting up an AI office in the UN Secretariat under which there will be several wings/initiatives with specific responsibilities. This includes (i) an independent international scientific panel on AI informing stakeholders with scientific research, (ii) a periodic intergovernmental and multi-stakeholder policy dialogue on AI governance, and (iii) an AI standards exchange formulating and maintaining AI standards.
Apart from having such global forums and institutions, AI Safety Institutes will also play an important role in global coordination on AI. Several countries like the UK, the US, and Japan have set up AI Safety institutes. These institutes can play an important role in fostering global cooperation and coordination on AI safety and AI governance. Consultations on establishing an AI Safety Institute in India are ongoing. Setting up such an institute in India will be key in ensuring that India’s perspectives relating to responsible and safe AI are well represented and included in global AI frameworks and standards.
You may contact the author at omir@cerai.in.