By Michael Yao, Sidwell Friends School
Artificial Intelligence is an incredibly powerful tool that has garnered a somewhat infamous reputation for its role in the targeted ads and eerily specific content recommendations of the modern world. However, AI’s aptitude at solving complex problems and rapid adaptation can also be used in more positive and impactful ways as well. One area that has attracted particular interest is its potential role to help resolve the climate crisis. In fact, research into how AI can be harnessed to address climate change has already begun.
Dr. David Rolnick (Assistant Professor of Computer Science at McGill University) and his work have been at the forefront of this push. Together with several other researchers, Rolnick co-authored a paper investigating how machine learning can address problems that other solutions may be less-suited or efficient at. In his paper, Rolnick identifies several areas that are of particular interest, dubbed “High-Leverage.” In an interview with Dr. Rolnick, I gained more insight into the particularities of these solutions, as well as other ways AI could be used to address climate change.
One way AI could help is by solving the problem of the “Duck Curve.” Put simply, the Duck Curve is a problem of supply and demand. Most commonly, it affects solar power, but is also a known problem for energy generation, notably wind. The problem occurs because solar energy generation tends to peak in the afternoon, but at the same time, demand for energy usually falls or stays even. This creates a problem of overproduction, where the solar panels are generating more energy than can be used. One solution may be to reduce energy production from other sources, but in many cases, power grid operators need those other sources to be running constantly to be profitable. This leads to the creation of an “artificial floor.” If energy from solar plus the “floor” passes demand, the operator has to turn off solar panels, wasting energy and valuable profits. In addition, it is difficult to rapidly ramp up production when solar production falters at night time, creating an additional issue. The same problem can be applied to other types of variable energy sources.
AI could help by doing what is called “now-casting”, says Rolnick, where AI performs “short term forecasting of solar and wind production by understanding how much the sun is shining or the wind is blowing, minute to minute.” In addition, AI could help create better charts of demand. “Right now [forecasting] is done by people.” This was the case because manual forecasting used to be more reliable and trusted by people over computer generated models. “But increasingly,” says Rolnick, “AI is getting to the point where it can often be trusted” to generate extremely accurate models.
Another high-leverage use of AI in shipping routes. “AI can help in selecting suppliers so that the overall emissions produced from shipping go down”, says Rolnick. Also, Rolnick says, AI could help in “picking suppliers, where shipments should come from, helping different modes of transportation interface with each other, like predicting train arrival times to help them run more efficiently.” Ultimately, AI’s impact on shipping is mainly one of optimization, helping various complex parts work in tandem and thus more efficiently, reducing the energy needed for work to be completed overall.
This idea of optimization is at the core of AI’s impact on the climate movement. Take the discovery of new materials, for example. Finding new materials for purposes of furthering sustainable technology is slow and inefficient, but AI could remedy this by recommending new materials to try based on existing data. While this method has no guarantee of success, it can help drastically reduce the amount of time and resources spent on discovering new materials by providing informed suggestions rather than using guesswork. In addition, AI could help lower the barrier for electric vehicle adoption, says Rolnick, “by helping companies design batteries that last longer”, and optimizing the energy expenditure of those batteries, thus increasing the car’s mileage per charge. Also, AI could help manufacturers decide where to build vehicle-charging stations.
The final area that Rolnick highlights is in the field of agriculture. AI could help farmers manage the growth of multiple crops at the same time, avoiding or mitigating the risk of stripping the farmland of its nutrients. “This is called precision agriculture, “ says Rolnick, “where there is very targeted robotics and sensors that are used to manage crops individually.” There are several options for implementing this. One way is by managing several different crops at once, which tends to be healthier than growing a single crop, otherwise known as monoculture. Farmers could also grow crops with trees or other plants to avoid monocultures. Additionally, says Rolnick, “there could be targeted applications of fertilizer and herbicide to specific plants in a field, so that less is used.” Another use for AI in agriculture is in the prediction of droughts, diseases, pests, and other problems that occur in nature. “This would be on a much smaller scale,” says Rolnick, “think months rather than years.”
Overall, the potential for AI to help solve climate change is exciting. However, Dr. Rolnick makes it clear that while problem solving can be done by machines, the issue is one that allows little time to solve, so the most impactful solutions to climate change will ultimately involve people, not robots or computers.
Work Cited:
“Electricity Production from Hydroelectric Sources (% of Total).” The World Bank, 2019, https://data.worldbank.org/indicator/eg.elc.hyro.zs?end=2019&start=2012.
Hao, Karen. “Here Are 10 Ways AI Could Help Fight Climate Change.” MIT Technology Review, 20 June 2019, https://www.technologyreview.com/2019/06/20/134864/ai-climate-change-machine-learning/.
Jones-Albertus, Becca. “Confronting the Duck Curve: How to Address Over-Generation of Solar Energy.” Office of Energy Efficiency & Renewable Energy, 12 Oct. 2017, https://www.energy.gov/eere/articles/confronting-duck-curve-how-address-over-generation-solar-energy.
“Renewables – Fuels & Technologies.” Iea, https://www.iea.org/fuels-and-technologies/renewables. Accessed 11 June 2021.
Vox. The “Duck Curve” Is Solar Energy’s Greatest Challenge. 2018. YouTube, https://www.youtube.com/watch?v=YYLzss58CLs.
About Author
