Governing artificial intelligence for planetary health

Area of Science:

• AI Governance
• Global Public Goods
• Risk Classification

Background:

• Increasingly pressing challenge to establish global AI governance.
• Diverse debates on AI focus on economic, social, environmental, or safety aspects.
• Need for coordinated interventions to mitigate AI's harmful effects.

Purpose of the Study:

• Classify AI risks and challenges across social, planetary, and safety domains.
• Propose AI governance as a global commons.
• Identify key regulatory dimensions for AI.

Main Methods:

• Risk and challenge classification across three domains.
• Analysis of inter-domain feedback loops and root drivers.
• Identification of regulatory dimensions: data, energy, and compute.

Main Results:

• AI risks are categorized across social, planetary, and safety domains.
• AI requires governance as a global commons with cross-domain interventions.
• Data, energy, and compute are identified as critical regulatory dimensions.

Conclusions:

• Emphasize limiting agentic AI and incentivizing depolarizing algorithms.
• Advocate for setting AI dynamics within global equity.
• Urgent need for coordinated global AI governance.