The increasing popularity of generative AI technology could worsen the ongoing e-waste problem by creating an additional 5 million metric tons of e-waste in total by 2030, researchers have said.

Generative AI works by using complex algorithms to create new content – such as text, images, music or video – that mimics or extends patterns found in existing data. But the technology is highly dependent on rapid improvements in hardware infrastructure and chipmaking that incentivises firms to undertake rapid hardware upgrades.

As well as the need to simply increase the computational power of the machines used to train the AI models, these models are also very power hungry, which comes at great expense. Moving to more energy efficient hardware can help to save energy costs in the long run.

A modeling study published in Nature Computational Science suggests that e-waste from the generative AI sector could increase from 2.6 thousand tons per year in 2023 to an estimated 2.5 million tons per year...