What is the Impact of AI on the Climate?


Artificial intelligence (AI) has become main subject  in our society, just like the climate emergency. 

While many are enthusiastic about the advances that AI will generate, including freeing humans from repetitive and mechanical tasks,others are concerned about the repercussions on jobs that it will surely transform or even eradicate. Kristalina Georgieva,Managing Director of the IMF (International Monetary Fund), estimates that nearly 40% of global jobs will be impacted and that, whatever the scenario, AI risks exacerbating inequalities.

More and more voices are also rising to denounce the ecological impact of artificial intelligence. The energy consumption necessary to train deep learning models, in particular, requires a lot of computing power. It is difficult to quantify its real footprint, but let’s take an example. The training of LLM GPT-3 by OpenAI generated the equivalent of approximately 500 tonnes of carbon dioxide.

Beyond greenhouse gas (GHG) emissions, AI can contribute to worsening already harmful activities such as optimizing mining operations – and therefore the production of barrels of oil – in the fossil fuel industry. AI also feeds recommendation algorithms that push for excessive consumption of fast fashion on social networks, while we know that the fashion industry already contributes 8% of global GHG emissions. And the examples are numerous in all sectors.

It could help reduce GHG emissions by 5 to 10% by 2030 according to the Boston Consulting Group, the equivalent of the annual emissions of the EU!

According to the World Meteorological Organization (WMO) of the United Nations, AI-based technologies offer unprecedented capabilities for processing and analyzing huge volumes of data, extracting relevant knowledge, and improving predictive models.

Artificial Intelligence allows:

  • Precise monitoring and calculation of the carbon footprint in all the complexity of the different elements that compose it
  • Optimized management of renewable energy sources thanks to better predictions of demand, allowing for more efficient consumption but also reducing waste
  • Less GHG-emitting precision agriculture, thanks to fine weather predictions to optimize water resource consumption but also reduce the use of pesticides and other fertilizers, while improving yields
  • Better anticipation of risks and adaptation to extreme weather events, allowing for more effective preparation of populations around the world to climate disasters
  • The implementation of more effective low-carbon policies
  • Optimization of transport or networks through flow optimization
  • The introduction of new generation electric vehicles with embedded AI
  • And we hope, the rationalization of the emission calculation process for all companies so that they can obtain more precise data on their ESG footprint, thus allowing them to quickly identify levers to reduce their emissions

However, for AI to be at the service of sustainable development and for new technologies as a whole to be more responsible and ethical; it is essential that governments commit:

  • By setting an example, with the use of ethical, ecological and low-carbon AI already in public services.
  • By legislation to regulate AI, with more transparency on algorithms and the development of more resource-efficient models.
  • By international cooperation to establish shared standards in the face of the global climate challenge.
  • Through incentives, such as carbon taxation for non-green data centers.

At Harington Impact, we work alongside you to put technological advancements at the service of a more sustainable and resilient future for our planet.

France Titin-Snaider, Associate President, Harington Impact

Other sources 

Learn more