Accepted Paper

Detecting Greenwashing on Social Media with CLAIMS (Climate Language and Image Monitoring System)  
Ned Westwood (Climate Accountability Lab, University of Miami) William Kattrup (University of Miami)

Contribution short abstract

The Climate Discourse Observatory at University of Miami’s Climate Accountability Lab have developed CLAIMS (Climate Language and Influence Monitoring System): a tool that automatically detects greenwashing in oil & gas companies' social media posts at scale.

Contribution long abstract

There is a growing need for tools that can detect false and misleading claims by fossil fuel companies in order to help inform climate accountability efforts such as lawsuits, advertising complaints, political investigations, and grassroots organizing. Based at the University of Miami’s Climate Accountability Lab, the Climate Discourse Observatory has developed CLAIMS (Climate Language and Influence Monitoring System): a tool that automatically detects greenwashing in oil & gas companies' social media posts at scale. By prompt-tuning the GPT5 Large Language Model from OpenAI, CLAIMS reliably classifies text in social media posts from fossil fuel producers according to a novel typology of “green” and “fossil fuel” messaging. Comparing the results to the real-world operations of carbon majors, we report the first quantitative evidence of greenwashing by fossil fuel companies on social media. In collaboration with the Algorithmic Transparency Institute, we have built an accompanying interactive CLAIMS dashboard that yields real time insights that may be relevant to ongoing accountability initiatives.

Roundtable P094
Corporate interference and false solutions - the Fossil Fuel Industry's obstruction in the energy transition