Overview

As AI is becoming ever more ubiquitous in society, many search engines now use generative AI to summarize search results in response to user queries, individuals can interact with generative AI chatbots such as ChatGPT to have it answer questions for them and media outlets increasingly turn to AI-assisted news production in the hope of enhancing journalistic efficiency and quality.

An important question emerging from these developments is whether AI-generated news stories have a greater impact on the public than human written news stories. While recent work has compared the persuasiveness of AI-generated versus human written public health messages, the genre of news stories, which most individuals come across on a daily basis and influences their political attitudes, has not yet been investigated.

For the topic of environmental change, in particular, it will be key to examine the perceived effectiveness of emotional framing in AI-generated news stories compared to those written by humans, as the way in which news stories are framed emotionally has been shown to have a strong impact on their persuasiveness. For this purpose, this project seeks to address four main research questions:

1. How persuasive are AI-generated news stories compared to news stories written by humans in encouraging individuals to take climate action themselves?
2. How persuasive are positively framed news stories compared to negatively framed news stories in encouraging individuals to take climate action themselves?
3. How do the effects of emotional framing on news story persuasiveness compare between AI-generated and human-written climate news stories?
4. What are the linguistic features of the news stories that persuade individuals the most to take climate action themselves?