Environmentally conscious AI
While the increasing use of AI has well-known negative environmental implications, our research found that AI also holds the potential to play a significant role in uncovering environmental damage or negative trends, and work towards addressing solutions for environmental sustainability.
We looked for initiatives that incorporate sustainable, environmentally conscious AI, which is limited, but is beginning to grow.
A notable example we came across as part of our research is the collaboration between Vodafone, Safaricom, and the World Wide Fund for Nature (WWF)in Kenya, which has developed an AI and Internet of Things (IoT)-powered system to reduce human-wildlife conflict.
This tool, m-Twiga (named after the Swahili word for giraffe), uses IoT devices and AI to monitor wildlife movement in Kenya’s Mara Siana Community Conservancy, and to alert rangers via SMS when wildlife is moving towards local communities, and activating deterrents like flashing lights and sounds to encourage wildlife to move to other areas.
We find that m-Twiga is a great example of how AI can address environmental and community challenges simultaneously.
Another example we came across as part of this research is the use of AI to highlight climate change. Senegalese researcher David Sagna developed machine learning models to predict rainfall patterns in Senegal’s Casamance region, a region particularly vulnerable to changes in precipitation.
Using K-nearest neighbor models (KNN), support vector machine (SVM), extreme machine learning (ELM) and predictive regression, Sagna’s work demonstrates how AI can help communities respond to climate challenges using past environmental behavior to predict future climate implications on communities.
In a world where climate change is impacting communities more and more, research like Sangra’s is critical towards anticipating possible next steps.
Last updated