It took about 40 years for humans, well scientists, to get any grasp on our Climate Change problem. And one of the MANY problems with climate change is we don’t have time.

 So this is where Artificial Intelligence could help us out.

At the moment, there are two different types to AI:

Rules-based and learning-based.

Rules-based AI need hard numbers to solve problems using if-then algorithms – good for helping scientists to crunch numbers, compile data and generally saving time.

However what we ideally need is predictions to the future. This is where Learning AI comes in. It has a memory, unlike rules-based AI, meaning learning-based AI can solve problems by interacting with them.

A simple scenario helps explain and differentiate the two AIs:

“Let’s say you asked a rules-based AI for a shirt. That AI would find you a shirt in the right size and color, but only if you told it your size and preferences. If you asked a learning AI for a shirt, it would assess all of the previous shirt purchases you’ve made over the past year, then find you the perfect shirt for the current season…”

smashing magazine

A few companies using Learning AI to help combat Climate Change.

SilviaTerra
Forests are important for our climate. The carbon dioxide that’s emitted by many human activities is actually absorbed by trees.

SilviaTerra uses AI and satellite imaging to predict the sizes, species, and health of forest trees.

DeepMind
Energy conservation through efficiency.

Google teamed up with the DeepMind AI to figure out how to save energy in their data centres. With the AI in place and learning how to use the least amount of energy Google cut their energy consumption by 35%.

Green Horizon Project
Measuring impact.

IBM’s Green Horizon Project is an AI that creates self-configuring weather and pollution forecasts. Between 2012 and 2017, Green Horizon helped the city of Beijing decrease their average smog levels by 35%.

CycleGANs
Extreme weather event predictions and outcomes.

Cornell University used GANs (Generative Adversarial Network) to make a self learning AI to output images of locations before and after extreme weather events. The graphics are used to help predict climate change impact and help us prioritise our efforts.