By Reese Meister
The recent growth and development of artificial intelligence (AI) has brought widespread attention to countless issues regarding copyright, employment, security and much more. However, some may not be aware of the environmental impacts that AI creates as well.
AI systems rely on immense amounts of energy and natural resources to function by powering supercomputers that process information. These systems need this abnormally large amount of energy because of the depth of information that they receive.
They often obtain electricity from the public grid with assistance from diesel generators, both of which can emit large amounts of greenhouse gasses into the environment. According to the Council of Foreign Relations, companies may use more than 250,000 pounds of carbon dioxide to train one AI system.
AI companies may deplete water sources in order to cool the supercomputers, which can reach high temperatures. Additionally, these computers contain chips that require especially high quantities of energy to produce.
Once the electronic devices and equipment have served their purposes, companies dispose of them as e-waste. Companies do not dispose of the majority of this e-waste in an environmentally sustainable manner, often leaving it in landfills rather than recovering the materials for reuse.
Most of these negative environmental consequences from the establishment of AI systems normally harm developing regions most severely, as issues such as pollution and extreme weather events are often already prevalent.
To combat these impacts, some groups have invested time in researching possible solutions that would increase the sustainability of AI systems. For example, the United Nations (UN) created the CODES Action Plan for a Sustainable Planet in the Digital Age in order to reduce AI’s emissions.
Participants on a Stanford panel, which addressed the environmental implications of AI, proposed solutions ranging from using smaller models to becoming more intentional and efficient in data collection. Relying on clean energy for larger projects and powering the system at times with lower energy demands could also help.
More sustainable power sources to consider could include solar energy, hydroelectricity or geothermal energy, a common form found in Iceland. However, geothermal energy will not be viable for every community; the best clean energy alternative depends on the resources of a region.
At the same time, companies have begun to use AI for environmentally beneficial purposes. A panel from the Organization for Economic Cooperation and Development mentioned technology’s “huge potential” to making a positive impact on the environment.
Among many other uses, AI can collect and analyze environmental data, which may help scientists develop ways to combat climate change and other forms of degradation. People can take advantage of this data to track anything from deforestation to ecological footprints.
AI’s algorithms provide data with improved accuracy, speed and effectiveness, which can assist governments and other groups who create environmental policy. The UN, for instance, made the World Environment Situation Room in 2022 to develop images of the earth and make predictions for the future. This program can track many types of environmental indicators to expand the UN’s access to information about the state of the planet.
The International Methane Emissions Observatory uses AI to combat emissions of this potent greenhouse gas through gathering information and producing global public records that aid scientists in making more sustainable decisions. Additionally, AI can allow mining companies to more efficiently find target minerals used to produce electric vehicle batteries, reducing the ecological harm from mining.
In more general examples, people may use AI for environmental design and sustainable development, improved satellite images and more. Self-driving cars that include AI have the potential to cut carbon emissions in half by 2050, and smart houses can also preserve energy and resources.
Since AI has only recently experienced such immense growth in a global spotlight, its systems lack regulations and guidelines to limit their environmental impact. Scientists and companies will continue to investigate methods of reducing the ecological impact of AI in order to enhance its ability to bring positive change to the planet.