Using a Large Language Model to Choose Effective Climate Change Messages

Using a Large Language Model to Choose Effective Climate Change Messages

- 1 min

Research carried at the ALFA group at the CSAIL, MIT. Authors: Thomas Benchetrit, Iris Kremer, Erik Hemberg, Aruna Sankaranarayanan, Una-May O’Reilly Presented at the PubLLM AAAI’24 Workshop


Using a Large Language Model to Choose Effective Climate Change Messages

Overview

“Using a Large Language Model to Choose Effective Climate Change Messages” delves into the capabilities of Large Language Models (LLMs) in shaping climate change communication strategies. The study investigates the model’s ability to create profiles, match them with specific audience groups, and select tailored climate change messages based on survey data and prompts.

Goal

We explore the potential of leveraging LLMs in crafting targeted communication strategies for climate change. We investigate the effectiveness of using LLM-generated tailored messages to enhance public understanding and engagement with climate change issues.

Methodology

Using structured prompts, the model generates profiles from survey data and matches them to pre-defined attitude groups (Profile Matching). Additionally, it selects suitable messages for these profiles based on their inferred preferences (Message Selection). The effectiveness of the LLM is evaluated by comparing its outputs to ground-truth data from established surveys and messaging techniques.

Interest

It emphasizes the potential of LLMs to extend traditional communication strategies by creating structured and targeted messaging techniques based on individual profiles. In particular, it bases itself on the previous Climate Change communication strategies of “Global Warming’s Six Americas,” which categorizes individuals based on their climate change attitudes, and explores the potential of LLM to correctly profile and provide tailored messages.

Applications

By demonstrating the adaptability of LLMs in crafting personalized messaging, the thesis contributes to understanding how AI can be leveraged to redefine public engagement on climate issues. It also paves the way for the potential integration of LLMs in environmental communication, offering new avenues for fostering public understanding and action on climate change.

comments powered by Disqus
rss facebook twitter github gitlab youtube mail spotify lastfm instagram linkedin google google-plus pinterest medium vimeo stackoverflow reddit quora quora