Generative AI Boom Could Trigger Millions of Tons in E-Waste, New Research Shows

According to a new study at Chinese Academy of Sciences’ Institute of Urban Environment , generative AI is going to produce million tons of e-waste in the upcoming year. The researchers said that they tried to add every device like batteries, circuit boards and other hardware that is being used to power generative AI applications in their study.

We all know that ChatGPT has taken over the world but there is a lot of hardware associated with it that is used to run the models. GPUs are used to run these applications, mostly in data centers and server farms. Generative AIs need energy and resources to work and they are also run on batteries in case of power outages because they are important for many users. But all the hardware and power equipment has a limited time under which it can work. After the time is over, this hardware can become e-waste.

The researchers studied the e-waste produced by AI applications and tried to estimate how much e-waste can be generated until the end of this decade. The researchers estimated the hardware which is used by AI applications and the data centers which are responsible for running many AI models. They also predicted how much the demand for AI applications is going to increase in the upcoming years.

The researchers said that if everything goes according to their estimates, AI is going to produce 1.2 to 5 million metric tons of e-waste. Right now, AI is producing 2.6 thousand metric tons of e-waste but it can increase to 2.5 million by the end of the decade. The researchers suggested recycling hardware if AI companies want to avoid such amounts of e-waste.

Image: DIW-Aigen

Read next: AI Job Growth Accelerates: Top Industries Seeking Skilled Professionals
by Arooj Ahmed via Digital Information World