Artificial Intelligence threatens to be a large contributor to climate change

Nidha Eakambaram

October 17, 2024

As AI grows more complex, the number of parameters on programs has increased (some up to 540 billion), which are run and stored in data centers. According to Columbia Climate School, at Columbia University, the data centers that run AI programs account for 2.5% to 3.7% of yearly global greenhouse gas emissions.

A survey by the Pew Research Center found that 55% of Americans said they regularly use AI. However, the environmental impact of AI is a growing concern due to the massive energy consumption required to train and operate these models.

Artificial Intelligence (AI) has brought transformative capabilities across various industries, from healthcare to logistics. It’s also transforming education by offering personalized learning tools like OpenAI’s ChatGPT to assist students with research and writing.

As reported by the International Energy Agency (IEA) in 2021, data centers hosting AI systems consume about 1% of global electricity, with numbers expected to rise as AI becomes more prevalent. According to the IEA, a majority of these centers still rely on fossil fuels, which creates a negative environmental impact.

A 2019 study by the University of Massachusetts Amherst found that training a single AI model (such as ChatGPT) can emit as much carbon dioxide as five gasoline-powered cars over their lifetimes or around 284 tons of CO2.

Junior Caroline Bingold is a leader of the Environmental Justice club at Lincoln and says there isn’t enough awareness of AI’s environmental effects.

“I don’t think most people know that things like ChatGPT have negative impacts on the environment,” she said. “I’ve never heard of AI polluting the environment.”

Junior Isa Mercado is also a member of the Environmental Justice Club (EJC) and is passionate about spreading awareness on this topic now that she knows about it.

“I think there’s a possibility that if people knew about how AI directly impacts people it might discourage them from using it,” she said.

AI developers and tech companies are exploring ways to mitigate the environmental effects of their technologies.

According to OpenAI, there are also ongoing efforts to make models like ChatGPT more efficient. By refining algorithms and optimizing the training process, developers hope to reduce the computational load and, in turn, the energy consumption required to run these models.

“I think for now being informed helps, and just being mindful of things and limiting use [of AI] is important,” Mercado said.