Below you will find resources that were released during the AI Impact Summit in India that took place from Monday, February 16 - Friday, February 20 or at least ones that I found that were released during it (lol).
This paper analyzes the environmental costs of AI systems and finds that while these impacts are growing, regulatory transparency has declined, showing that most global regulations operate at the facility rather than model level, missing the vast differences in energy consumption between AI types. To address this gap, the authors propose mandatory model-level transparency requirements, user rights to opt out of unnecessary AI, and international coordination, with concrete legislative amendments for the EU and other jurisdictions.
This paper argues that AI governance requires a global commons approach addressing interconnected social, planetary, and safety risks, identifying data, energy, and compute as critical regulatory dimensions that span all three domains. The authors emphasize the need for coordinated interventions that tackle root drivers like monopolistic AI power and polarizing algorithms, while prioritizing global equity and limiting highly autonomous systems.
This analysis challenges tech companies' claims that AI will solve climate change, arguing that their promises lack substantiated evidence while AI's current environmental costs, particularly data center energy demands are proven and growing, effectively giving the industry a blank check to pollute based on speculative future benefits.
In recent years, the time to connect a data center to the power grid has grown significantly. In response, many data center developers are choosing to build their own power plants and avoid the grid or connect later. In what we believe is the most comprehensive analysis of this trend to date, we identified 46 data centers with a combined capacity of 56 GW that plan to build their own power "behind-the-meter." That represents roughly 30% of all planned data center capacity in the United States, according to Cleanview's project tracker.
The objective of this approach is to ensure efficient use of resources, reduce confusion, promote consistency in the measurement of the environmental impact of Artificial Intelligence (AI), and facilitate the widespread adoption of best practices in that regard. Contributors wish to work towards non-conflictual standards to measure the environmental impact of AI and encourage collaboration between international standardization bodies to avoid, as far as possible, the duplication and overlaps of standards. Version 2 was prepared for AI Impact Summit in India.
Artificial intelligence (AI) is set to transform economies and societies worldwide, with significant implications for people and the planet. For developing nations, AI will bring both transformative benefits and risks, requiring a proactive approach to its regulation that builds safeguards while promoting innovation. This report therefore provides an assessment of the potential multidimensional impacts of AI on the people and countries of the global South, particularly on their digital transformation, labour and industrial development.