By: Anurag Jakkula
Climate change has caused glaciers to shrink, animal migrations to shift, trees to flower sooner, and ecosystems to diminish. The fact is, climate change has been the result of our technology, and now it is up to our technology to handle it. Artificial intelligence, or AI, is the science of simulating human intelligence with technology. Machine learning is one aspect of AI which is proving to be a powerful tool in the battle against climate change. Put simply, a machine learning algorithm does not react to an input. Rather, the machine “learns” the algorithm and “finds” the output, improving its algorithm with new data. Machine learning shows promise in monitoring climate change, spreading awareness, enforcing emissions, and the technology to reduce emissions.
AI plays an integral role in the research of climate change. In 2011, a discipline at the intersection of climate science and data science was created. Machine learning allows for never-before precision in predicting and observing climate change. For example, a study by Ise and Oba harnessed the power of deep learning, which is a subcategory of machine learning, and considered the most advanced aspect of AI. Deep learning draws its inspiration from the human brain, as it mimics a neural network. The study fed global temperatures from the past thirty years to a neural network. With no other data, the algorithm predicts the temperature fluctuations for the next ten years, with a 97% accuracy. To enact positive change, people need to understand that climate change is a reality. However, awareness can only be spread with the availability of information. AI can provide us with this information.
After the information and predictions are present, a simulation which predicts the appearance of homes after being affected by climate change related factors, such as rising sea levels and intense storms, has been created by researchers from Montreal Institute for Learning Algorithms (MILA), Microsoft, and ConscientAI Labs. A P.h.D. candidate at MILA and co-author of the project’s research paper, Victor Schmidt, expresses, “Our goal is not to convince people climate change is real, it’s to get people who do believe it is real to do more about that.” In order to spread this awareness and call for action, the project’s team plans to release an app which shows the simulation. In addition, people will be able to upload pictures of forest fires, floods, storms, and other climate change effects to improve the algorithm.
Climate change is caused by the emission of fossil fuels. AI can allow for the enforcement of emission limits for coal plants and factories. Carbon Tracker, a London-based non-profit, monitors coal-plant emissions by using satellite imagery. Due to a grant from Google, Carbon Tracker is working to expand its efforts to include gas powered plants. All in all, around 4000 to 5000 power plants will be observed and analyzed, creating the biggest public data bank. AI has the potential to automate the analysis of these images to pinpoint problematic entities. If a carbon tax is passed in the future, Carbon Tracker’s efforts could help enforce the tax.
Lastly, AI plays a huge role in technology that directly limits emissions. For example, Google’s DeepMind, an artificial intelligence company based in the UK, implemented a machine learning algorithm to optimize the energy consumed by Google Data Centers, which currently accounts for 3% of the world’s energy consumption. By optimizing the settings and cooling these data centers, the algorithm could reduce the energy consumption by 40%. In addition, Microsoft uses AI-driven underwater data centers. The data centers’ AI algorithm allows complete self-sufficiency, with wave energy and ocean cooling.
There is no question as to whether AI could make a tremendous positive impact in the issue of climate change. However, AI is not the solution by itself. Rather, it is a powerful component that needs to be incorporated with other disciplines in order to succeed against such a huge issue. MIT, Carnegie Mellon, Google, and other big names in AI have published a paper called, “Tackling Climate Change with Machine Learning.” The paper identifies thirteen areas where machine learning could make a major impact, including energy production, CO2 removal, education, solar geoengineering, and finance. A collaboration between climate specialists, engineers, government, entrepreneurs, and AI scientists is of the utmost importance, and the effort looks promising.
What Did You Learn?
1. What is machine learning and deep learning?
Machine learning is an aspect of AI. In machine learning, the machine does not take inputs in order to emit an output. Rather, the machine “learns” the algorithm in order to come up with the output. The algorithm improves itself with the new data. Deep learning is a subfield in machine learning. A deep learning algorithm emulates the neural network of human’s brain. It is considered the most advanced aspect of AI.
2. What role does AI play in reducing emissions?
AI can help data centers, which use a lot of electricity, to become more efficient. For example, the DeepMind machine learning system in Google’s data centers could reduce their energy consumption by 40%. Microsoft has self-sufficient underwater data centers that rely on wave energy and are cooled by the ocean. The Microsoft data centers are driven by AI.