Artificial intelligence (AI) and digital tools are increasingly being regarded as a way to foresee and reduce its effects as climate change accelerates the destruction from storms, wildfires, and droughts. Machine-based learning systems, which employ algorithms to find patterns in data sets and make predictions, suggestions, or judgments in real-world or virtual contexts, are gaining interest from governments, tech companies, and investors. In order to predict weather patterns with fine accuracy, Kentucky-based startup Climavision developed a data- and AI-driven “nowcasting” system.
In June, the Rise Fund, an impact investing division of private equity firm TPG, invested $100 million in this method. Also, at the COP26 climate summit in Scotland in November, an intergovernmental roadmap on AI’s role in combating global warming is scheduled to be unveiled. But, skeptics point out that AI can also be very energy-intensive and environmentally harmful, and they caution that the technology might be an expensive diversion from more efficient strategies to combat climate change.
How can AI help combat climate change?
The technology is already in use to create greener smart cities in China, monitor deforestation in the Amazon, and deliver natural catastrophe alerts to Japan. Moreover, AI applications could optimize the deployment of renewable energy sources by supplying solar and wind energy into the electrical grid as needed, increase power storage, and design more energy-efficient structures. On a lesser scale, it might assist families in reducing their energy consumption by automatically turning off lights when not in use or rerouting electricity from electric cars to the grid to fulfill expected demand.
According to a recent research conducted by the accounting company PwC for Microsoft, which is creating machine learning solutions for the climate change sector, the technology might help reduce global greenhouse gas emissions and air pollution by 4% by the year 2030. Co-founder of the British think tank Centre for AI and Climate (CAIC), Peter Clutton-Brock, believes the technology is “pushing back barriers” for climate modeling. He thinks AI opens “great opportunities for understanding the dynamics underlying sea level rise and ice sheets” by processing vast volumes of unstructured data like images, graphs, and maps.
Who will be able to use it?
The technology is already in use to create greener, smarter cities in China and to transmit natural disaster alerts to Japan, track deforestation in the Amazon, and monitor it. By supplying solar and wind energy into the grid when needed, AI applications might also optimize the deployment of renewable energy sources by designing more energy-efficient structures, enhancing power storage, and designing buildings with better energy efficiency.
On a smaller scale, it might aid families in reducing their energy consumption by automatically turning off lights when not in use or feeding power from electric vehicles back into the grid to fulfill expected demand. According to recent research conducted by the accounting company PricewaterhouseCoopers for Microsoft, which is developing machine learning solutions for the climate change sector, found that the technology might help reduce global greenhouse gas emissions by 4% by 2030. Technology is “pushing back barriers” for climate modeling, according to Peter Clutton-Brock, co-founder of the think tank the Centre for AI and Climate (CAIC), based in Britain. He told the Thomson Reuters Foundation that AI offers “great opportunities for understanding the dynamics of sea level rise and ice sheets” by processing vast volumes of unstructured data such as images, graphs, and maps.
What’s the downside?
According to a Massachusetts Institute of Technology study, the data storage and processing required to completely train a complex algorithm can use a significant amount of energy, producing up to 626,000 pounds (284,000 kg) of carbon dioxide.
That is equivalent to nearly five times the lifecycle emissions of an American car. The International Energy Agency estimates that the processing and storing of data from internet activities like sending emails and streaming films already makes up 1% of the world’s electricity use.
Furthermore, according to some projections, the global energy requirement for computers will reach 8% by 2030, which raises concerns that additional fossil fuels will need to be used as a result. Professor of social and ethical AI at Umea University in Sweden, Virginia Dignum, asserted that “AI is both a facilitator and, potentially, a destroyer of the climate fight.” She also noted that AI itself is “not a magic wand – and not without mistakes,” adding that the necessity for rare earth metals to produce the hardware has a detrimental effect.
When it comes to the possibility to “backtrace data to individuals,” as Clutton-Brock at the CAIC put it, the possible logging of people’s energy use raises privacy concerns. Depending on the assumptions employed as data inputs, biased predicted results are potentially a possibility. According to Kevin Anderson, a professor of energy and climate change at universities in Britain, Sweden, and Norway, addressing climate change should “mainly (be) about those of us who are responsible for the lion’s share of emissions making significant changes to our lives.”
So, what are the solutions?
In the end, how AI is governed internationally may determine how it is used to combat climate change. Dignum, a member of the High-Level Expert Group on Artificial Intelligence of the European Commission last year, said that privacy concerns are “particularly peculiar to the type of technology that we are employing at this time.”
“The trade-off between energy usage and privacy becomes less of a challenge if we add other sorts of algorithms that are less strongly dependent on personal data,” she said. These algorithms would soon be “as efficient” as those in use now, she continued. Green AI principles have been presented by the AI Now Institute, a research center at New York University, and include the integration of tech and climate legislation as well as complete transparency regarding a tech company’s carbon, energy, and environmental effect. The European Union has hinted that it is also thinking about these problems. Yet for the time being, Clutton-Brock added, efforts to use AI to discover a solution will continue due to the risks posed by climate change.
