In recent years, technology and investments in generative artificial intelligence (AI), such as ChatGPT, Midjourney, and other large language model (LLM)-based technologies, have surged dramatically. This has led to a range of accompanying benefits and drawbacks, as noted by IFLScience.
While there are numerous debates surrounding misinformation and existential fears, researchers caution that there is one far more practical issue that has gone largely unnoticed: an astonishing amount of electronic waste is accumulating on the planet.
According to Asaf Tzachor, a sustainable development and climate researcher at Reichman University in Israel, electronic waste generated by AI, particularly large language models, is expected to increase sharply by 2030. Scientists predict it could reach 2.5 million tons per year by the end of this decade unless immediate action is taken.
Tzachor co-authored a new study examining the potential consequences of electronic waste resulting from the sudden boom in generative AI technologies. The researcher and his colleagues found that the volume of electronic waste from generative AI computer servers could reach an astounding 5 million tons per year by the decade's end. For comparison, this is nearly 2000 times the amount of waste produced in 2023.
According to the United Nations, electronic waste is already a significant issue for the planet today. For instance, in 2022, 62 million tons of electronic waste were generated, and scientists note that this would fill 1.55 million 40-ton trucks — enough vehicles to circle the Earth's equator.
The authors of the study emphasize that addressing this problem will not only prevent the death of millions due to toxic chemicals and pollution but will also provide the world with an additional resource — electronic waste is overflowing with valuable minerals and resources. For example, there is more gold and silver in a ton of iPhones than in a ton of ore from a gold or silver mine.
Researchers found that implementing circular economy strategies could reduce electronic waste generation by 86%. In the worst-case scenario, waste could still be reduced by 16%. This study makes it clear that the electronic waste crisis is global in nature, making it crucial to focus on transboundary electronic waste management.
The authors of the study note that there is no universal solution to this problem; however, Tzachor and his colleagues identified several strategies: