We need a race to develop AI for good: to develop AI that… supercharge climate action [ and ] an AI that propels us towards the Sustainable Development Goals.
Today, the climate crisis is so dire that a rush to find any solution that “supercharges” climate action is inevitable. The latest hysteria of artificial intelligence has propelled it as the go-to tech for solving all the world problems, including climate change. Though AI could transmogrify as a potent solution for all climate problems, one should not discount its environmental costs. To address this dichotomy, a balance, weighing the help and the hurt that AI can be for the climate, has to be achieved.
How Can AI Help in Climate Action?
Addressing the climate issue rests on comprehending the historical trends of the change, analyzing its causes, and coming up with actionable insights. It would mean gleaning vast amounts of complex data — from cameras to sensors to satellites — and making sense out of it. Here is where AI shines.
Predict Future Climate Patterns
AI models are extensively used today to predict future climate patterns. Regional modeling of sea level rise, wildfires, floods, and droughts helps forecast extreme weather events by location. Insights from the AI systems would give policymakers data-driven awareness of rising sea levels and impending weather hazards and help develop accurate early warning systems to alert communities. They can better plan infrastructure investments to construct buildings resistant to natural disasters.
Predict Supply of Renewables
Machine learning models are also accurate in telling in advance how the supply of renewables, such as wind and solar energy, will be the next day. By combining the historical energy production data with current weather inputs and synthesizing missing data, these models provide timely and reliable renewable energy forecasts. Accurate predictions of wind or solar energy that would be available when given hours in advance, thanks to AI, would help authorities plan the energy supply. They are, hence, prepared to meet all the energy demands on a given day without downtime. Optimized grids go a long way in achieving Sustainable Development Goal (SDG) 7 – Ensure access to affordable, reliable, sustainable, and modern energy for all.
Optimize Carbon Offsets
Artificial intelligence also plays a role in optimizing carbon offsets. From the enormous volumes of data, it can uncover the impact of each carbon emitter and measure the carbon’s natural stock remotely. AI is also effective at a micro level as it can calculate the carbon footprint of individual products. The models could assess carbon capture storage sites and improve carbon sequestration practices. It can alert people of the dangerous levels of the polluted air in a particular area by monitoring urban air quality reports. Using AI in carbon markets would help set carbon prices precisely and backed by data. Furthermore, tracking air quality with AI would give insights to improve urban planning, public health, traffic, and waste management.
Improve Crop Yields
Furthermore, AI models could anticipate the best planting times by combining the insights of weather events, renewable energy forecasts, and climate effects. And hence improve crop yields. They could assess soil health and monitor pest and disease outbreaks. Intelligent irrigation systems powered by AI could reduce water usage at these times when droughts and famines are increasingly becoming both frequent and protracted. With improved modeling of climate change patterns, communities can plan effective strategies for better crop yield.
Optimize Supply Chains
Another area where AI benefits climate is by reducing industrial waste by optimizing supply chains, monitoring resource consumption, and planning recycling efforts.
Prevent Biodiversity Loss
Finally, the burgeoning AI market can help prevent biodiversity loss. Machine learning models help monitor forests and natural reserves’ encroachment. They are used to identify and periodically count various species. Consequently, we are able to predict their large scale migration patterns. Monitoring biodiverse lifeforms using AI predicts the probabilities of species extinction and hence gives an edge to policymakers to create effective strategies to preserve and restore the planet’s ecological diversity.
But Will AI Harm Our Environment?
As with any other technology, AI is not the panacea for all climate issues. In fact, this vaunted technology comes with its own baggage of climate problems.
The carbon footprint of AI models is lethal enough to often put discussions on using it for solving climate quandaries into a stalemate. Though the environmental harm depends on several factors, including model type, size, number of parameters used, amount of data processed, and data centers running these models, it is not uncommon to categorically associate AI models with harmful carbon emissions. The desire for frequent release of faster and more complex hardware to improve the model accuracy results in intensive cycling through the hardware, rendering a lot of hardware very useless very quickly. But, with hardware recycling still struggling to catch up, this e-waste induced by the booming AI industry is stressing the environment like never before.
Further divesting AI from becoming the answer to climate trouble is the condition of the existing solutions. Current AI-based climate change remedies are sparse and hard to access and scale. Besides, AI is used by the fossil fuel industry to accelerate oil and gas production and swell their coffers. Consequently, AI often receives desultory responses from climate proponents.
The Middle Ground – Dealing with the Dichotomy
Using AI to solve the climate crisis while minimizing its adverse effects on the environment could be done in a few ways. It could begin with contemplating whether AI is a must for the climate project — leave it off the tech stack if found unnecessary. Instead of developing a new model every time, it is prudent to use the existing models and tweak them to the underlying data and conditions. Smaller models are more energy efficient than their larger counterparts. Promoting the use of renewables for powering these models would further reduce the harm to the environment. Improving stakeholder engagement would contribute to better recycling practices. Access to capital and trained professionals could bridge the gap between academic research and at-scale deployments of fit-for-purpose and inclusive AI models. Instead of a fragmented approach, integrating AI for climate into national policies would help plan a comprehensive, global scale AI strategy for climate action. Scientific bodies, under the auspices of governments, should come up with standards for AI safety reporting. Individuals and companies should be held accountable for non-compliance.
In conclusion, despite the mixed impact machine learning has on climate change, using AI to address the climate crisis does have some merits. Artificial intelligence, if used right, can be a catalyst to achieve SDG goal 13 – Climate Action and combat climate change and its impacts.
