AI Tools Shape Efforts to Combat Air Pollution, Raise Energy Concerns
Image Credit: Maxim Tolchinskiy | Splash
Artificial intelligence is increasingly influencing strategies to address air pollution and climate change, offering new capabilities for data analysis and policy planning, while posing challenges tied to its own environmental impact. The World Health Organization (WHO) estimates that air pollution, particularly fine particulate matter known as PM2.5, contributes to roughly 7 million deaths each year, a figure that has shown little improvement in recent years.
[Read More: Smart Cities, Smarter Solutions: How AI is Revolutionizing Urban Life]
AI Enhances Pollution Monitoring and Policy
AI technologies are being applied to analyze vast datasets, refine climate models, and shorten the time needed for accurate projections, aiding efforts to craft targeted environmental policies. These tools can now estimate air pollution exposure at highly specific levels, such as individual neighbourhoods, offering detailed insights into how PM2.5 and similar pollutants affect populations.
This capability aligns with recent regulatory changes in the United States. On February 7, 2024, the Environmental Protection Agency (EPA) updated its National Ambient Air Quality Standards (NAAQS) for PM2.5, reducing the annual health-based limit from 12 to 9 micrograms per cubic meter. The adjustment aims to lower the incidence of heart attacks, premature deaths, and other health issues linked to particle pollution. The EPA’s regulatory impact analysis suggests that current regulations and available controls could cut PM2.5 levels, potentially delivering up to US$46 billion in public health benefits by 2032. AI is supporting these efforts by modelling the combined effects of reducing air pollution and greenhouse gas emissions from shared sources, including coal-fired power plants, diesel vehicles and industrial operations.
Historical data underscores the economic case for such measures. A 2011 EPA study projected that the Clean Air Act Amendments of 1990 would generate US$2 trillion in benefits by 2020, outweighing costs by more than 30-to-1. Globally, the nonprofit Clean Air Fund reports that air pollution currently results in 1.2 billion lost workdays annually, a number that could climb to 3.8 billion by 2060 without intervention. AI’s ability to process data is helping identify policies that could mitigate these losses while supporting economic productivity.
[Read More: Do you know the ESG aspect of AI?]
AI’s Energy Use Sparks Debate
Despite its contributions, AI’s environmental footprint is drawing attention. The energy required to power AI systems, particularly in data centers, is rising sharply as the technology expands. The International Energy Agency reported that in 2022, global data centers accounted for 1-1.3% of total electricity demand. In the U.S., a Harvard study, co-authored by an environmental researcher, found that carbon emissions from data centers tripled since 2018, reaching 105 million metric tons of CO2 in the 12 months ending August 2024. This represents 2.18% of national emissions, compared to 131 million metric tons from domestic commercial airlines.
Energy consumption by U.S. data centers has doubled since 2018, now accounting for 4.59% of the country’s total energy use. As AI becomes more embedded in environmental strategies, its emissions could offset some of the reductions it helps achieve.
Weighing Benefits Against Costs
Researchers suggest that innovations like more efficient AI algorithms or data center cooling methods, such as free air cooling, could reduce the technology’s carbon footprint. However, these approaches are still in early development, and their long-term impact remains uncertain. The task ahead is to maximize AI’s ability to identify at-risk populations and inform climate adaptation while managing its energy demands.
The evolution of solar panels provides a useful comparison. In the 1990s, early solar technologies, such as first-generation silicon panels, were criticized for their inefficiency, with energy payback times—the period needed to generate the energy used in production—ranging around 4-7 years for silicon-based panels, according to estimates from that era. High energy costs in mining, refining and manufacturing, often powered by fossil fuel-heavy grids, meant the net energy return on investment (EROI) was relatively low compared to modern standards.
However, technological advancements since then, including improved cell efficiency and cleaner production methods, have significantly reduced payback times. According to the National Renewable Energy Laboratory (NREL), modern solar panels can now achieve energy payback times between 0.5 and 1.2 years in optimal conditions. Over their 25-30 year lifespan, these panels generate 20 to 60 times the energy required for their production.
Similarly, AI’s current energy burden, driven by data centers, could diminish with efficiency gains, potentially amplifying its net benefit in pollution reduction—though, as with solar, the outcome hinges on how quickly and effectively those improvements are realized.
The urgency is clear when comparing standards. The WHO advises that annual PM2.5 levels should not exceed 5 micrograms per cubic meter, a threshold stricter than the U.S.’s new 9.0 limit. Remote areas with minimal pollution, recording levels as low as 1-3 micrograms per cubic meter, illustrate the potential for improvement.
[Read More: AI Kicks Away Bitcoin? The Power Struggle for Energy Resources]
Data Guides Dual Fight Against Pollution and Climate Change
The overlap between air pollution and climate change offers a strategic opportunity. Sources like fossil fuel industries and open burning contribute to both PM2.5 and greenhouse gases, meaning reductions in one area can benefit the other. AI’s capacity to analyze these connections is informing efforts to address both issues simultaneously. Between 1970 and 2019, U.S. emissions of common air pollutants dropped 77%, while GDP grew 285%, demonstrating that environmental gains need not hinder economic progress.
Regardless of debates over climate change’s origins, the health and economic toll of pollution and climate-related events—such as wildfires, heatwaves and storms—is well-documented. AI’s real-time data analysis is enabling tailored responses to these threats, though its energy-intensive nature complicates the picture.
[Read More: SkyVisor Secures €1.2M to Advance AI-Driven Drone Inspections in Renewable Energy]
Source: Forbes
