Spotting climate misinformation with AI requires expertly trained models

Meta's Llama and Google’s Gemini lag behind proprietary ones in the task

Illustration of a hand holding a smartphone engulfed in flames, symbolizing the spread of climate misinformation.

Organizations that want to counter climate misinformation with AI need to bring in experts to guide training of the models, a new study suggests.

rob dobi/Moment/Getty Images Plus

Conversational AI chatbots are making climate misinformation sound more credible, making it harder to distinguish falsehoods from real science. In response, climate experts are using some of the same tools to detect fake information online.