WNAM REPORT: The UN Climate Change Conference, COP29, began with the gathering of heads of state and government on the second day. The international event, attended by more than 100 world leaders, gathered on the second day to discuss more important issues and put forward initiatives.
There were also many experts from around the world participating in COP29. They exchanged views on important issues on side events and also shared their views with local and international media.
Observer and AI expert Shrey Goyal unpacks the difference between Sustainable AI (designing AI systems that are environmentally conscious) and AI for Sustainability (leveraging AI to reach sustainable development goals) in his interview. In this interview, we dive into the groundbreaking potential of Sustainable AI and how it can transform our approach to climate change.
We learn about the innovative ways AI is being used to reduce energy consumption in tech infrastructure, boost renewable energy use, and track environmental changes like deforestation. Plus, we discover the critical role of global collaboration—between governments, tech companies, and environmental organizations—to ensure AI’s development benefits both for society and the planet.
Explaining the concept, Goyal emphasized, “It’s essential to distinguish between sustainable AI and AI for sustainability. Sustainable AI is about ensuring that AI development and deployment are sustainable, adhering to environmental norms, and in some definitions, also ethical and social standards. Conversely, AI for sustainability uses AI to achieve sustainable development, environmental, social, and governance goals.”
Goyal highlighted how AI supports climate action: “AI can help reduce the carbon impact of ICT technologies, such as using smart cooling for energy-intensive servers. It can optimize energy systems by determining the most efficient times for solar or wind energy. Additionally, AI can track deforestation and enhance system efficiency across various processes.”
He discussed the technologies addressing AI’s carbon footprint, noting that “AI training has a huge carbon impact; training a medium-sized AI model consumes energy equivalent to five cars over their lifetimes. Reducing this is vital, especially with AI potentially consuming up to 20% of global electricity by 2030. New algorithms for efficient training and federated models help by reducing data transfer energy. Using renewable energy and placing data centers in cooler regions can further reduce energy consumption.”
On the necessity of collaboration, Goyal stressed, “Collaboration is crucial. AI’s rapid growth poses regulatory and societal challenges, affecting everything from environmental impact to job markets. International organizations like the UN and NGOs can drive collaborative efforts. For example, the UNFCCC’s technology mechanism includes an AI for Climate initiative with Google, Microsoft, and small island developing states, though more initiatives are needed.”
Reflecting on the role of the private and public sectors, he concluded, “Our organization is also aiding governments in AI use. The commercial sector often leads in adopting new technology, sometimes for purposes like marketing or data exploitation. It’s crucial to enable AI’s potential for the social sector and governments as well. AI holds great promise but also poses risks, so we need to proceed with caution.”
