February 14, 2024
3 minimum read
The echo chamber of climate change denial on social media is strongest in the Midwest and South, and in states that rely heavily on fossil fuels.
The lake bed of Nicasio Reservoir cracks during the drought on Wednesday, October 13, 2021 in Nicasio, California, USA.
Climate wire | Nearly 15 percent of Americans still deny climate change is happening, according to a new study that used artificial intelligence to collect data from social media.
This conclusion is similar to other recent findings. But by using deep learning models based on the technology used by ChatGPT, researchers collected data from his We were able to analyze how they are correlated.
The findings — published Wednesday in the journal scientific report — shows that people with similar climate views tend to group together on social media, forming echo chambers that reinforce their views. This makes it difficult for scientists to combat misinformation online, the authors suggest.
About supporting science journalism
If you enjoyed this article, please consider supporting our award-winning journalism. Currently subscribing. By subscribing, you help ensure future generations of impactful stories about the discoveries and ideas that shape the world today.
“What’s frightening, and somewhat discouraging, is how divided the world is between belief in and denial of climate change,” said the study’s lead author, a professor of environment and sustainability at the University of Michigan. Joshua Newell said in a statement. There is little communication or interaction between echo chambers. ”
The findings also show that some social media personas have significant influence within their respective silos. Former President Donald Trump was the most influential of the X accounts that denied the science of climate change.
The study collected information from X posts published by users in the United States between September 2017 and May 2019. The researchers used a deep learning model based on artificial intelligence company OpenAI’s GPT-2 language model to classify posts as either believing or denying their existence. climate change. Researchers tried to eliminate postings by bots by deleting the data of users who posted more than 20 times in a day.
The findings suggest that 14.8 percent of the U.S. public denies the reality that the planet is warming. These opinions tended to be highest in the south-central United States. It was also highly correlated with Republican voters and states that rely heavily on fossil fuels for energy.
The study found that climate change denial tends to be less correlated in communities with higher COVID-19 vaccination rates, and that rejection of basic climate science often goes hand in hand with other forms of scientific skepticism. This suggests that there is a possibility that
Echo chambers with different climate views had little interaction with each other and were less likely to repost or cite posts from communities with opposing beliefs. That means it can be difficult for scientists to engage with social media users who hold opposing views.
Still, the authors say their study allows experts to identify communities based on geography, politics, and other characteristics that are more likely to reject climate change science, allowing for more targeted strategies to combat disinformation. He said it could be useful for standing up.
This study had several limitations. In general, there are fewer contributions to means.
Additionally, because the language model classified only posts that believed in or denied global warming, the research method may have excluded people with neutral or undecided views on climate change.
Reprinted from E&E News with permission of POLITICO, LLC. Copyright 2023. E&E News provides news that matters to energy and environment professionals.
