Navigation
Search
|
The Impact of Artificial Intelligence on Pollution: Opportunities and Challenges
Thursday November 7, 2024. 11:26 PM , from eWeek
KEY TAKEAWAYS •Training and operating AI models consumes electricity, emits greenhouse gases, overuses water, creates e-waste, and puts pressure on natural resources. (Jump to Section) •AI technology can mitigate climate change and pollution by helping us track it in real time, predict it in the future, and drive innovations in anti-pollution technology. (Jump to Section) •Despite slow progress, governments are taking a bigger role in regulating AI development and incentivizing sustainable AI. (Jump to Section) TABLE OF CONTENTS ToggleUnderstanding the Impact of AI on PollutionHarnessing AI to Combat Artificial Intelligence PollutionGovernmental Regulatory Efforts: Addressing AI PollutionFrequently Asked Questions (FAQs)Bottom Line: Balancing AI’s Potential with Environmental Responsibility Understanding the Impact of AI on Pollution Artificial intelligence contributes to air, water, and earth pollution in three ways. First, producing AI models requires vast amounts of energy typically produced by burning fossil fuels. Second, IT equipment relies on non-renewable resources mined in environmentally damaging ways. Third, IT equipment at the end of its lifecycle can become e-waste that ends up in rivers, landfills, and oceans. AI’s Carbon Footprint: Energy Consumption and Emissions Training AI models, especially large language models like Chat-GPT, is an energy-intensive process that emits a massive amount of CO2 into the atmosphere. Umass Amherst researchers found that training an AI model emits five times as much carbon dioxide as fueling and manufacturing the average American car. The major culprits are the data centers where many AI models are trained and housed. Data centers require a staggering amount of energy to run complex AI algorithms and power their electronics. To put it in perspective, the average Google search electricity demand is about.3 Wh, while the average for a Chat-GPT query is 2.9 Wh—about 10 times higher. Data centers currently consume about 1 to 2 percent of global energy. Goldman Sachs predicts that by 2030, that will rise to 3 or 4 percent. Unfortunately, much of that energy still comes from burning fossil fuels, which releases greenhouse gases (GHG) into the environment. If data center companies like Amazon fail to find more energy-efficient methods and switch to more renewable energy sources, carbon emissions and pollution from AI will continue to rise and negatively impact climate change. Electronic Waste: The Hidden Cost of AI Technologies Greenhouse gas isn’t the only environmentally harmful byproduct of building and maintaining AI models. Data centers also create improperly discarded equipment, known as electronic waste or e-waste. Common examples of hardware that becomes e-waste include routers, circuit boards, power supply units, cooling systems, and network switches. These electronics often contain hazardous chemicals like lead and mercury, which degrade the environment by poisoning soil, wildlife, and bodies of water. This could lead to health issues among humans who live off these contaminated resources. The AI boom drives the construction of more data centers, which require more hardware—leading to more e-waste. Statista predicts that by 2030, 82 million metric tons of e-waste will be created each year. If the past is a predictor, much of that waste is unlikely to end up in recycling centers. In 2022, only 23 percent of the 62 million metric tons of e-waste was formally collected and recycled. Much of the remainder likely ended up in landfills. If the mismanagement of e-waste continues, data centers will remain a growing environmental concern. Impact on Natural Resources: Water Usage and Resource Depletion Many AI data centers rely on cooling techniques that require large amounts of water to prevent computers and IT equipment from overheating. One common method is evaporative cooling, where water is evaporated to chill the air around the equipment. Once hot, the air is released into the atmosphere. Systems like this often run continuously. The hotter the equipment gets, the more water is used. Unfortunately, AI workloads—combined with increasing heat due to climate change—are making data centers hotter than ever. The exact amount of water consumed by data centers is hard to calculate. NPR reports that the average mid-sized data center uses around 300,000 gallons of water per day. In comparison, the average person consumes around 100 gallons of water a day. Droughts already threaten millions of people around the world. Climate change, with its warmer temperatures, makes these droughts longer and more severe. It’s critical that AI data centers find more efficient ways to cool their IT equipment and waste less water. Water, however, is not the only resource that AI data centers gobble up. IT equipment like semiconductors, LEDs, and electronic boards are built with natural resources, including rare minerals and earth materials. Computer drive boards contain the highly rare elements dysprosium and yttrium. Other earth elements needed for data centers include copper, gold, silver, lithium, and silicon. These materials are often mined in unsustainable ways that can disrupt ecosystems, pollute the air, contaminate groundwater, and cause dramatic changes to the landscape through deforestation, sinkholes, and erosion. AI vendors are being pressured to be more transparent about their natural resource use, but the process has just begun. Harnessing AI to Combat Artificial Intelligence Pollution Although artificial intelligence contributes to pollution, the technology can also prevent and mitigate it by optimizing energy use, monitoring pollution and environmental harm, and powering new technologies that clean up pollution. Optimizing Energy Use with AI: Potential and Applications AI can help data centers, buildings, and businesses reduce energy expenditure by automatically analyzing energy usage data and identifying efficient ways to run. This AI-driven energy management could offset some energy loss from AI-related activities like cooling data centers and training AI models. For example, a data center could use an AI-powered air conditioning system that tracks room and computer temperatures, predicts future temperatures, and automatically adjusts the thermostat accordingly. This way, energy use is optimized. In other words, the air conditioning system is never on when it could be off, saving electricity. Some buildings, like The Edge in Amsterdam, use AI technology to create more energy than they consume. In addition to solar panels, The Edge uses IoT smart sensors across the building to monitor real-time data about occupancy, light levels, temperature, and humidity. This data enables their systems to manage heating, lighting, and ventilation effectively. Whether it’s a data center or an office building, AI-driven energy management can help them reduce their energy use and, as a result, lower their carbon footprint. A modern marvel of green engineering, The Edge creates more energy than it uses. AI in Environmental Monitoring: Tracking and Predicting Pollution Artificial intelligence has powerful use cases in environmental monitoring and pollution detection. Real-time monitoring facilities can identify pollution at any moment in the day, allowing organizations to intervene quickly to clean up the current contamination. Some technologies can even identify the source of the pollution, enabling agencies to catch the perpetrator, find the problem, and prevent it from happening again. AI also helps predict and prevent future pollution. By applying AI-powered predictive analytics to historical pollution and air quality data, urban planners can forecast how much a certain project will pollute the air. Accurate pollution tracking and forecasting help environmental organizations and governments make environmentally conscious decisions. Innovative AI Solutions for Reducing Ecological Impact For now, pollution is inevitable. As emerging economies industrialize, new smoke stacks will rise as forests fall. Even in the most modern, green economies, companies still rely heavily on burning fossil fuels for energy—and despite large-scale recycling projects, plastic still ends up in rivers. However, innovators are using AI to build technologies that can reduce the ecological impact of such activities. For example, The Ocean Cleanup leverages AI technology to find and remove ocean litter, protecting wildlife and reducing the GHG emissions from plastics. Another company, Space Intelligence, uses AI and satellite imagery to track deforestation rates. Companies like this can give environmental organizations the data they need to protect forests while allowing green-conscious investors to put their money into non-deforestation initiatives. From a climate change standpoint, intervention in deforestation is critical since tree burning releases a lot of CO2 into the environment. The Environmental Defense Fund states that about 20 percent of annual global greenhouse gas emissions come from tropical deforestation. Trees absorb CO2 out of the atmosphere, making them our strongest ally in the fight against climate change. If we incentivize and invest in companies that are “greenovating,” or innovating with green technologies, artificial intelligence can live up to its promise as a major supporter of the environment. Space intelligence combines machine learning, satellite imagery, and ecology expertise to identify and monitor deforestation. Governmental Regulatory Efforts: Addressing AI Pollution Government AI policy is a key aspect of reducing artificial intelligence pollution. Legal entities can create laws and policies that prompt AI-related companies to use less water, energy, and vital resources, and produce less e-waste and GHG emissions. Governments can also incentivize or force companies to disclose pollution, carbon emissions, and other environmentally harmful practices, ultimately increasing accountability in AI development and operations. Most countries have only set up guidelines for the sustainable development of AI, and few have created real legislation. However, governments seem to be moving in this direction. In the U.S., in 2024, 31 states adopted resolutions or enacted legislation for regulating AI. Most related to the safe use of AI, but some focused on AI and the environment. For example, Hawaii required the University of Hawaii to create a program for building an AI-powered wildlife forecast system to predict wildfires. New AI bills are introduced regularly. Congress recently introduced the Artificial Intelligence Environmental Impacts Act of 2024, which calls for the Environmental Protection Agency to study AI’s environmental impacts and develop a voluntary reporting system. Such legislation is a positive sign for the future that shows the government taking AI’s environmental impact seriously. As scientists discover more about the connection between AI and the environment, legislative efforts to address environmental concerns should increase and gain more traction. Frequently Asked Questions (FAQs) How Can AI Help with Ocean Pollution? AI can combat ocean pollution by analyzing satellite data to track plastic waste, optimizing cleanup vessel routes, monitoring ecosystems, locating pollution sources, and automating waste sorting. Predictive AI helps agencies and nonprofits predict pollution movement patterns for swift intervention. AI-powered vessels can also use cameras, robotic arms, and AI imaging to collect, identify, and sort ocean pollution, allowing for efficient recycling. Are Generative AI Apps Bad for the Environment? Generative AI apps can be harmful. AI art generators like Midjourney or Dall-E are trained and operated in data centers requiring massive energy. This energy is often generated by burning fossil fuels, which releases greenhouse gases into the atmosphere. These data centers often create e-waste (old IT equipment) that contains harmful chemicals that can damage the environment. Do Robots Pollute the Environment? Robots can contribute to environmental pollution throughout their lifecycle. During manufacturing, builders might use unsustainable practices that rely on fossil fuels. In operation, robots can use a lot of energy and electricity, which may also require fossil fuels. When disposed of, robots can contribute to e-waste, as some are made of environmentally damaging materials. On the bright side, robots can also help reduce pollution by increasing industrial efficiency and improving environmental monitoring and cleanup. Bottom Line: Balancing AI’s Potential with Environmental Responsibility Artificial intelligence is changing how humans work, opening opportunities for business innovation, and underpinning a new technological revolution. However, AI also raises serious environmental concerns. AI development is an energy-intensive process that contributes to pollution in various ways, from producing e-waste to using fossil fuels to power the data centers that train AI models. These data centers also use enormous amounts of water and natural resources, the mining of which can harm the land. AI businesses must find more efficient and sustainable ways to develop AI, and governments should incentivize these innovations. Otherwise, artificial intelligence risks being just another technology that pollutes our environment, damaging the ecosystems, wildlife, and human lives that depend on it for survival. Read AI for Climate Change to learn how this emerging technology can benefit the environment. The post The Impact of Artificial Intelligence on Pollution: Opportunities and Challenges appeared first on eWEEK.
https://www.eweek.com/artificial-intelligence/artificial-intelligence-pollution/
Related News |
25 sources
Current Date
Nov, Fri 8 - 04:38 CET
|