The “Other Side” of AI Systems

Artificial intelligence systems (AI systems) are being used more and more often all over the planet. We usually think that AI systems are making our lives better or more comfortable. What most people do not know is that using AI systems has significant negative impacts on the environment and ecosystems, particularly where Indigenous People or other local communities are living.

Think about AI chatbot. Say you want to build a small gingerbread house as a Christmas present for your mother, but you do not know how to do it. Instead of looking for a book where that is explained, you ask an AI chatbot. Within seconds, you have different options on how to create a gingerbread house. This seems simple, but the answers AI chatbot gave you are the result of a complicated and vast global supply chain that can damage the environment.

How AI Systems are Built

Do you know how AI systems are built and what they need to work properly? Generally, when people think of AI systems, they think of something “in the cloud” that cannot be touched. In reality, AI systems require various devices, like microchips, to work properly. Microchips, are made of raw materials that must be extracted (removed) from the Earth by mining in different parts of the world. AI systems also need special buildings, called data centers, designed to hold many powerful computers called servers. A server is a big computer that helps other computers talk to each other and share information. Finally, once microchips and other electronic components are old or broken, they are thrown away as electronic waste (e-waste), polluting the environment. Let us look at these three problems in a bit more detail.

Extracting Raw Materials

To manufacture the microchips and other electronic equipment necessary for AI systems, we need to dig (extract) the raw minerals out of the ground. Numerous mining sites are located in forested areas, especially in rainforest areas like the Amazon, which are cut down to dig for underground minerals. Mining is considered the fourth main cause of deforestation, and computer and electronic devices are responsible for 5% of all deforestation worldwide.

As a consequence of deforestation, animals can lose homes, soil can wash away. In addition, water can be contaminated with dangerous chemicals and released into the surrounding areas. This can cause problems for people and nature [1]. People living close to mining areas often do not have clean water to drink and their environments may be polluted, making them unhealthy. For example, local communities can suffer severe diseases due to exposure to toxic substances released into the soil, water, or air near their homes [2]. People working in the mines often face dangerous working conditions. Another consequence of mining-related pollution is the displacement of Indigenous Peoples and local communities living in the surrounding areas, which violates their fundamental rights.

According to experts’ predictions, the demand for minerals needed to manufacture the components of AI systems, such as lithium or cobalt, will surge alarmingly by 2050. The consequences of such a growing demand could be extremely negative for fragile ecosystems, biodiversity, and local communities.

Data Centers

To function properly, AI systems must be trained by analyzing lots of data on how to recognize, for example, a dog. When many pictures of dogs are shown to an AI system, the system learns what a dog looks like. AI systems are trained on powerful servers located in data centers (Figure 1). Data centers consume huge amounts of energy and of water—the largest data centers consume approximately 1%-2% of the electricity used throughout the world. The servers at data centers get really hot and must be cooled down with water. For example, in 2022, Google’s data centers used approximately 20 million liters of mostly drinkable water to cool servers. Experts have predicted that the global AI demand will account for 4.2–6.6 billion cubic meters of water use in 2027 alone.

Furthermore, people can be exploited for the training of AI systems. Workers that provide the labeled data that AI systems need to learn are often paid very little, frequently not enough to cover basic needs like food, rent, and clothes, and they are not legally protected [3].

E-waste Disposal

When an electronic product is no longer useful it is thrown away as e-waste. E-waste is growing at an alarming rate. In 2022, 62 billion kg of e-waste was generated worldwide. Experts predict that this amount will increase to 82 billion kg by 2030. E-waste is usually dumped in big trash piles called landfills, many of which are found in poorer countries far away from where most AI devices are used [4]. In those countries, children and adults who need money often sort e-waste by hand and can get sick because of chemicals like mercury or arsenic. These chemicals can also seep into the water and soil, polluting the surrounding environment, poisoning these natural resources and harming plants, animals, and people [5].

Sometimes e-waste is burnt by people looking for shiny metals such as gold. Burning plastic and wires creates toxic fumes that hurt the lungs of people (including children) working or living nearby, and the gases released also contribute to global warming. In addition, workers who recover materials and valuable parts of e-waste can experience poor working conditions, and over 18 million children from 5–17 years are exploited by industries in the e-waste sector.

What are the Possible Solutions?

To sum up, AI systems are contributing to deforestation all around the world; polluting the air, soil, and water with toxic materials; and consuming huge amounts of electricity and water. In addition, people living close to mining areas or data centers, or those working in mining sites or in e-waste landfills can get sick from toxic chemicals that seep into environment. Children exposed to these chemicals may have more trouble learning, playing, and growing up strong and healthy. Understanding the ecological footprint of AI systems is thus extremely important so that we can use these technologies with more awareness of their negative impact.

So, what can we do about it? First, raising awareness about the ecological footprint and social consequences of AI systems is extremely important. Many people are concerned by these impacts, especially children and teenagers, who will be most affected by the negative effects of climate change. Yet, children and teenagers are also using AI systems more and more, maybe without knowing that they have such negative consequences.

Second, it is important to use AI systems more responsibly. Experts and lawmakers are working on solutions, from cooling data centers underwater to recycling e-waste. But responsible use of AI systems and education are some of the most important tools. You can be an advocate for responsible AI use and inform others about the negative impacts of AI systems. Even small steps, like using AI systems only when they are really necessary, can make a big difference and inspire others to so the same. If enough people act, we can keep the benefits of AI without destroying the only planet we have. The future of the Earth is in our hands.

Glossary

Artificial Intelligence Systems (AI Systems): ↑ Various techniques enabling machines to simulate human intelligence.

Global Supply Chain: ↑ The journey needed for a product, such as an AI tool, to exist. For AI, it includes steps from mineral extraction to manufacturing, shipping, data storage, and e-waste disposal.

Raw Materials: ↑ Natural things that have not been changed or processed by humans, from which a product is made.

Data Centers: ↑ special buildings designed to hold many powerful computers called servers used to store, process or distribute a large amount of data.

Servers: ↑ Big computers that help other computers talk to each other and share information.

E-waste: ↑ old or broken electrical and electronic equipment, such as smartphones or micro-components of electronic devices that contain a mixture of valuable materials, such as gold, and toxic materials.

Fundamental Rights: ↑ Basic rules ensuring that all individuals are safe, healthy, and respected.

Ecological Footprint: ↑ Degree of damage and destruction of the environment caused by human activities.

Conflict of Interest

The author(s) declared that this work was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.

Acknowledgments

This study was funded by the European Union. Views and opinions expressed are however those of the author only and do not necessarily reflect those of the European Union or the granting authority, i.e., European Research Executive Agency (REA) under the powers delegated by the European Commission. Neither the European Union nor the granting authority can be held responsible for them. Project “EcoAI - A New EU Framework for an Ecological AI”, HORIZON-MSCA-PF-2023, Marie Sklodowska Curie Actions – Post doctoral Fellowships – Global Fellowships, GA 101155739.

AI Tool Statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.