Artificial intelligence boom could generate 5 million tons of e-waste by 2030: study warns

A recent study conducted by environmentalists at the Chinese Academy of Sciences in collaboration with Reichman University in Israel raises serious concerns about the environmental impact of generative artificial intelligence (AI).

According to the study, demand for specialized hardware needed to run applications like ChatGPT is growing rapidly, rendering older electronic devices obsolete.

“Our results suggest that this e-waste stream could increase, potentially reaching a total accumulation of 1.2 to 5 million tons between 2020 and 2030,” the researchers said. South China Morning Post reported.

The boom in artificial intelligence will generate approximately 2,600 tons of e-waste in 2023 alone. However, according to the study, if preventive measures are not taken, this number could increase to almost 2.5 million tons by 2030.

To put this into perspective, the amount of e-waste generated would be equivalent to each of the approximately 8.5 billion people on the planet throwing away almost two smartphones.

The price of the artificial intelligence boom on the planet

The study highlights how rapid advances in artificial intelligence are making old equipment obsolete and how constant replacement is exacerbating the problem of e-waste, which poses serious risks due to the presence of toxic materials such as lead and chromium.

Meanwhile, current e-waste management procedures often have to catch up, even though many of these components contain recyclable metals such as gold and platinum.

According to the study, most of the e-waste will come from North America, particularly the US and Canada, where many artificial intelligence data centers operate.

The study argues that geopolitical restrictions, such as the US ban on selling sophisticated GPUs to China, could also worsen the situation. Chinese data centers may end up using outdated technology, which could lead to inefficiencies and more e-waste.

Such a strategy “could lead to a loss of computing efficiency in countries that do not have access to such chips, leading to increased demand for physical servers,” the researchers say.

The study also compares Nvidiaa modified export version of the H800 that has less ability to comply with export laws to China, with the latest H100 chip, which US regulators have banned from being sold to Chinese consumers due to national security concerns.

“For example, the Nvidia H800’s bandwidth efficiency is half that of H100which requires doubling the number to achieve equivalent performance,” the researchers noted.

“Our analysis shows that a one-year delay in receiving the latest chips could lead to a 14 percent increase in end-of-life e-waste, totaling 5.7 million tonnes between 2023 and 2030, more than the global total. small electronic waste. information and communication technology waste in 2022.”

A personalized e-waste management strategy can help.

Researchers support a circular economy strategy that emphasizes recycling and reusing electronic components to address this looming disaster.

This tactic could reduce e-waste production worldwide by up to 86%.

Wang Peng, the lead author, stressed that different regions should take individual approaches to solve their specific problems.

“The US should focus on more responsible hardware manufacturing and algorithm development as it is at the forefront of artificial intelligence development. The global artificial intelligence industry could become more resilient if its sources were managed properly,” Peng said, according to SKMP.

“For China, the focus will be on strengthening regulation of exploitation and disposal processes. AI hardware is produced in large quantities and then frequently upgraded, eliminated, and replaced. And regulation of waste management must be tightened.”

The study was first published in Nature Computational Science.