AI Can Help Lessen Dangers of Climate Change, Industry Leaders Say

A study co-written by leading academics and business leaders analyzes nine different ways artificial intelligence can help reduce carbon emissions

Almost all scientists agree that human behavior is causing climate change, and now some academics and business leaders say it will take superhumans to help fix it.

A recent study by academics at several universities in collaboration with researchers at Google and Microsoft analyzed nine ways artificial intelligence (AI) could help reduce carbon emissions that are harmful to the Earth's atmosphere. It offered four tools to take action on the proposals.

Global leaders and citizens can use machine learning, a section of AI that focuses on computational data analysis, to create more efficient electricity systems and transportation models, to predict food and weather forecasts, and, eventually, to remove carbon dioxide from the atmosphere, the research paper said.

Some of the technology suggested, such as drones and computer-monitoring of behavior, is not new. In fact, drones were being pitched to farmers to improve agriculture as long as eight years ago, said Marc Conte, an associate professor of economics at Fordham University, in a phone interview. Conte was not involved in the study. 

What the study calls for is more incorporation of the technology in society, and lays out ways of doing so.

It also stresses that AI is not the only solution needed to solve the problems brought by climate change, but can be one of the significant pieces to help, said David Rolnick, the study’s lead author and a mathematical sciences postdoctoral research fellow at the University of Pennsylvania.

“Climate change does not present one problem, it presents multiple problems. AI is only one of the tools that can have an impact in the fight to mitigate the effects of climate change,” Rolnick said.

The proposals for climate mitigation outlined in the study require collaboration, Rolnick said. For data to be drawn together to make conclusions, everyone must share their information.

“We can use these tools together with existing tools to help all stakeholders,” Rolnick said. “Above all we have to start working together.”

As has happened in the past, people will undoubtedly raise security concerns about the data that is being collected and who is collecting it, Conte said.

Despite this, he still sees the overall benefits of the all-encompassing technology.

“It’s hard to argue that as we increase dramatically our access to computation power, that machine learning is not going to play a large role in improving our lives,” Conte said.

Here are the ways the study says AI can make our world greener and more efficient:

Electricity Systems: Electricity that relies on fossil fuels currently makes up 25% of global carbon emissions. AI can ensure that researchers and developers know of developments in clean electricity technologies, such as solar, wind, hydro and nuclear. AI can make better models to forecast electricity use and improve system management for homes and buildings. Behavior can be monitored so less fuel will be used when it is known that the residents will not be there.

Transportation Systems: Transportation emissions from cars, trains and planes account for about a quarter of the world’s total CO2 emissions. AI can help increase vehicle efficiency by improving engine design and aerodynamics, and by managing logistics of cleaner power sources, for example locating power stations for electric vehicles with a batteries.

Buildings & Cities: Also making up 25% of the world’s carbon emissions, buildings can be fixed with easy-to-implement solutions, according to the study. Emissions can be cut with more efficient heating and cooling systems that study residents' habits. 

Industry: AI analysis can help make more efficient supply chains. The advanced technology can also use past data to predict machine breakdowns.

Farms & Forests: Plants and algae naturally release carbon dioxide into the air as part of photosynthesis. AI can be used to institute precision agriculture, a practice that tries to predict what crops will be in demand at the beginning of the season, to reduce carbon released from the soil, help people monitor forests and predict forest fires.

Carbon Dioxide Removal: As one of the more long-term goals, removing carbon dioxide from the atmosphere with the help of technology seems to be far off from wide-spread use. Yet using AI will help find good locations with facilities designed to store CO2 most efficiently.

Adaptation: Climate change predictions are based on models that the study's authors believe can be improved. AI can make climate models more accurate, predicting more exact future temperature increases due to carbon emissions happening today. It can also create long-term models to predict extreme weather forecasts such as hurricanes.

Societal Impacts: Lower crop yields and local food shortages may become a greater problem in the future. In building systems to predict food shortages, AI can manage historical data and documented trends to identify areas with the highest risk of shortages and can get food to areas that need it.

Solar Geoengineering: Releasing certain emissions into the atmosphere, if the right kind, can have the opposite effect of CO2 emissions that cause climate change. Engineering substances released in the atmosphere to act as a shield, blocking sunlight and global warming, is one way the crisis can be mitigated. Gases can also be manipulated to help heat escape.

Presented with predictions and models that AI can provide, individuals and businesses will better understand their carbon footprint. They then can make decisions that avoid climate change's negative effects and change their behavior, the report concluded. 

“We will have a smaller or bigger crisis depending on what we do now,” Rolnick said. “We will have a crisis. But we can change the nature of that crisis.”

Contact Us