Skip to content

The high cost of Artificial Intelligence: energy consumption and tons of pollution

It’s fashionable. Many people want to use it, either for research, curiosity, or play. The Artificial intelligence (AI) seems to be everywhere. However, making this technology work has a high cost that can harm the environment.

Powering AI requires millions of processors, graphics cards, hard drives, and RAM working in unison. According to research carried out by Bloomberg, this produces tons of polluting gases resulting from stratospheric consumption.

READ ALSO: Counter-Strike 2: this is how the legendary Dust II, Overpass and Nuke maps look in the new game

The work you do can take days, months, and years collecting information. And it is that the models use large doses of cloud computing (cloud computing), indicates the page El Español.

Now, according to research, each AI model consumes more power than any other computing system, and training them consumes the equivalent of 100 US homes in a year.

Another investigation carried out by the University of California at Berkeley, in 2021, points to some estimates on the energy consumption of ChatGPT-3 and that it needed 1,287 GW/h, which is equivalent to the consumption of 120 average American homes.

Despite this, OpenAI has indicated that it works to reduce the impact on the environment and in a reduction in both energy consumption. “We take our responsibility to stop and reverse climate change very seriously,” the company said.

Source: Elcomercio

Share this article:
globalhappenings news.jpg
most popular