SCIENCE

AI Reveals Hotspots of Climate Denial

[ad_1]

Echo chambers of climate denial on social media are strongest in the U.S. Midwest and South and in states that depend heavily on fossil fuels

A cracked lake bed with mountain in background.

A cracked lake bed at Nicasio Reservoir during a drought in Nicasio, California, U.S., on Wednesday, Oct. 13, 2021.

CLIMATEWIRE | Nearly 15 percent of Americans still deny that climate change is happening, according to a new study that used artificial intelligence to gather data from social media.

That conclusion is similar to findings from other recent surveys. But by using a deep learning model — based on the technology used in ChatGPT — researchers were able to gather data from X, formerly known as Twitter, and analyze how climate science denial correlates with other factors.

The findings — published Wednesday in the journal Scientific Reports — show that people with similar climate views tend to group together on social media, forming echo chambers that reinforce their own views. That makes it harder for scientists to combat misinformation online, the authors suggest.


On supporting science journalism

If you’re enjoying this article, consider supporting our award-winning journalism by subscribing. By purchasing a subscription you are helping to ensure the future of impactful stories about the discoveries and ideas shaping our world today.


“What is scary, and somewhat disheartening, is how divided the worlds are between climate change belief and denial,” said senior study author Joshua Newell, a professor of environment and sustainability at the University of Michigan, in a statement. “The respective X echo chambers have little communication and interaction between them.”

The findings also demonstrate that some social media personas have an outsized influence within their own silos. Former President Donald Trump had the biggest reach among X accounts denying the science of climate change.

The study collected information from X posts published by U.S. users between September 2017 and May 2019. Researchers used a deep learning model, based on the artificial intelligence company OpenAI’s GPT-2 language model, to classify the posts as either believing or denying the existence of climate change. Researchers attempted to screen out posts by bots by removing data by users who posted more than 20 times a day.

The findings suggest that 14.8 percent of the U.S. public rejects the reality that the Earth is warming. These views tended to be highest in the central and southern United States. They were also highly correlated with Republican voters and with states that heavily depend on fossil fuels for energy.

Climate denial also tended to have a low correlation with communities that had high Covid-19 vaccination rates, suggesting that the rejection of basic climate science may often go hand-in-hand with other forms of scientific skepticism, according to the study.

Echo chambers with different climate views tended to have little interaction with one another and low rates of reposting or quoting posts from communities with the opposite beliefs. That means it may be difficult for scientists to engage with social media users who hold opposing views.

Still, the authors said the study may help experts develop more targeted strategies for combating disinformation by identifying communities — by geography, politics or other characteristics — that are more likely to reject the science of climate change.

The study had some limitations. There are generally fewer posts on X from geographic areas with smaller populations, including rural communities, and the smaller sample size means there is greater uncertainty about the connections between these locations and their climate views.

The study’s methods may also have left out people who have neutral or undecided views on climate change, as the language model only classified posts that believed or denied that the Earth is warming.

Reprinted from E&E News with permission from POLITICO, LLC. Copyright 2023. E&E News provides essential news for energy and environment professionals.

[ad_2]

Source link

Related Articles

Back to top button