Why AI Infrastructure and Governance Must Evolve Together
The rapid deployment of artificial intelligence (AI) technologies is transforming various sectors, including data infrastructure and global governance. However, the evolution of AI infrastructure is outpacing the regulatory frameworks that are needed to govern it effectively.
AI Infrastructure Challenges
As AI models grow in size and complexity, their physical footprint, including compute power and thermal management, is surging. This increase raises significant concerns regarding energy consumption, reliance on non-renewable resources, and the production of e-waste.
The Jevons Paradox suggests that efficiency gains in resource usage might actually lead to an overall increase in resource consumption as the usage of AI technologies rises. This paradox highlights the necessity for a sustainable approach to AI development and deployment.
The Governance Lag
Governance frameworks, including environmental laws and digital regulations, are struggling to keep pace with the rapid advancements in AI infrastructure. This disparity creates a new tension: AI infrastructure is evolving faster than the regulations needed to ensure it serves the public interest and planetary health.
This mismatch influences policymaking, business strategies, and infrastructure investments. Addressing these gaps is crucial for achieving a balance between technological advancement and the sustainability of our resources.
Critical Mismatches in AI Governance
Despite the apparent separation of infrastructure and digital governance, these domains are converging. However, this convergence remains underexplored in public discourse. The current landscape reveals three critical mismatches:
- Functional mismatches: Silos exist among AI infrastructure, environmental sustainability, and sectors such as finance.
- Spatial mismatches: There is insufficient coordination across local, national, and international governance scales.
- Temporal mismatches: The rapid deployment cycles of AI systems clash with the long-term needs of environmental and societal resilience.
To address these mismatches, a holistic approach is required, where infrastructure development and governance evolve in tandem. This is particularly critical in regions like the Asia-Pacific, where urban density and climate vulnerability intersect with accelerated digitalization.
Singapore’s Initiatives
Singapore exemplifies proactive governance through its Green Data Centre Roadmap and the Model AI Governance Framework, which aim to align infrastructure and governance frameworks for sustainable digital ecosystems.
The Infrastructure Shift
AI development is increasing the demand for advanced digital infrastructure, particularly data centres. This ecosystem faces stress in various areas, including electricity supply and e-waste management.
Traditional air cooling systems are reaching their thermodynamic limits, making the transition to liquid cooling technologies crucial. These solutions improve thermal transfer efficiency and reduce the energy and spatial footprints of data centres.
The Need for Comprehensive Environmental Assessments
A comprehensive environmental assessment of AI infrastructure should consider Scope 1, 2, and 3 emissions, aligning with the Greenhouse Gas Protocol. This includes evaluating energy usage, embodied carbon in manufacturing, and material impacts at the end of the lifecycle.
Regulatory Innovation
Regulatory frameworks must adapt to the rapid development of AI infrastructure. Without clear, interoperable sustainability standards, there is a risk of fragmented compliance, regulatory arbitrage, and inconsistent sustainability reporting.
Clear standards should support environmental impact assessments and cross-border regulatory coherence. As AI systems become more complex, governance frameworks must remain flexible to accommodate these variations while maintaining sustainability goals.
A Cooperative Process for Sustainable AI
Achieving sustainable AI requires integrated actions from governments, industry players, academic institutions, and the broader community. Cooperative frameworks are beginning to emerge globally, with the European Union’s AI Act 2024 mandating energy consumption documentation for general-purpose AI models.
Asia has a unique opportunity to lead in sustainable AI governance by adopting best practices and shaping future standards. Key areas for prioritization include:
- Conducting full lifecycle impact assessments.
- Managing infrastructure complexity.
- Improving transparency in resource usage.
- Fostering regulatory innovation.
- Advancing cross-border standardization.
By directly linking infrastructure development with environmental responsibility, regions can demonstrate what integrated, forward-looking AI ecosystems can achieve.
