A productive artificial intelligence (AI) system can consume up to 33 times more energy for certain tasks than machines running software. The information comes from Dr. Hugging Face, a company specializing in machine learning. It comes from a study conducted by Sasha Luccioni.

Many generative AI systems are trained with large repositories of knowledge to generate answers to various types of queries. Each query results in information being created from scratch, which requires a lot of computing power. Dr. According to Luccioni, this entire process is extremely inefficient.

It is worth noting that currently most of this computing is performed not on our computers, but in huge data centers dedicated to productive AI tasks, among others.

In 2022, data centers consumed 460 terawatt hours of electricity worldwide. The International Energy Agency (IEA) expects this number to double in just four years.

It’s hard to predict how much energy AI will consume

Currently, some experts are warning that there may soon be an explosion in the energy consumption of AI data centers, Dr. He agrees with Luccioni’s concerns. In the UK alone, this consumption could increase sixfold in just 10 years, according to National Grid.

But there are also technologists who alleviate the problem; According to them, the evolution of chips will be responsible for keeping energy consumption at acceptable levels.

Some experts say that the energy efficiency of new artificial intelligence chips will be enough to alleviate energy consumption.

There are several factors that can affect the global electrical energy demand for AI processing. Many chipmakers such as Intel, AMD, Apple and Qualcomm are already introducing components that can process AI tasks directly on users’ computers, reducing the demand for cloud processing.

In this case, energy consumption passes from data centers to users’ homes. It is not yet known to what extent the energy efficiency of enterprise and end-user chips can contain the harmful growth in energy consumption.

Meanwhile, Dr. Luccioni is working on a project to improve energy ratings for AI to facilitate the adoption of more efficient systems.

Source: Tec Mundo

Previous articleRussian scientists have found a way to extend the service life of solar panelsScience and technology21:46 | May 26, 2024
Next articleState e-sports center opened in St. Petersburg
I am a passionate and hardworking journalist with an eye for detail. I specialize in the field of news reporting, and have been writing for Gadget Onus, a renowned online news site, since 2019. As the author of their Hot News section, I’m proud to be at the forefront of today’s headlines and current affairs.

LEAVE A REPLY

Please enter your comment!
Please enter your name here