Teamname: Cecece Slogan: Rethinking Climate Communication Team members: J├Âran Landschoff, Li Cheng, Theresa Fink, Torsten Landschoff

The problem

Climate change communication is often not very effective to reach new and less involved audiences, however, it is very important to generate and communicate and unite a large societal basis for actions and policies to combat climate change and its negative effects. Issues posing difficulties for this communication are the abstract nature and slow processes of climate change that are not easily recognizable by the individual. Additionally, societies are strongly fragmented with varying degrees of prior knowledge and interest in climate change leading to polarized attitudes concerning beliefs about climate change. While communication specifically targeted at subgroups can be effective, social media posts can be seen by a large audience including less involved or antagonistic and diverging groups. Climate change communication on social media should, thus, consider the effects of posts on less involved groups and avoid further polarization caused by exclusive or aggressive language.

Our solution

To aid climate change communicators, we developed the vision for a software tool that identifies problematic language in social media posts and suggests possible improvements. With this tool, we want to reduce further polarization of opinions on climate change and lead to a more inclusive language through which people that did not identify with the movement feel respected and invited to join in action against climate change. Posts are evaluated based on the tone and sentence structure and polarizing language of "us" versus "them" is then suggested to be replaced alongside an explanation for that reason. While social media is frequently used to communicate with supporters it is read by and affects other, often unknown people. Our tool checks posts considering this and encourages the use of more inclusive and motivating language that can win over more supporters.

What we did

During this weekend, we developed the project idea and started analyzing social media posts of climate change communicators, such as Fridays for Future, with respect to possible negative impressions on groups not yet involved in the movement. We found alternative solutions and phrases that refrain from blaming people that are needed for successful action and contain more inclusive and motivating language. Our aim is to automate this identification of problematic language and the suggestion of improved versions with an AI-based algorithm.

What's next

In the future, this algorithm has to be developed. The characteristics of problematic language have to be classified and pinpointed further. Categories have to be identified that can then be used to "flag" posts to indicate to the user to revise the text of the post. A training set of marked social media posts of problematic sentence structures and semantics has to be generated to train the algorithm for detecting unwanted text patterns. Additionally, the algorithm has to learn how to propose suggestions for improved examples, which would require a second training set with positive examples.

Built With

  • none
Share this project: