AI-generated content is increasing on the Internet – including in journalism. But a study has now revealed that many people view AI-generated headlines rather skeptically.
Since the introduction of ChatGPT, the topic of artificial intelligence has become an integral part of many areas of everyday life. The new technology does not stop at news journalism either.
This means that more and more AI-generated texts, images and even videos are being created, some of which are also used in the journalistic field. But how do consumers react to this AI-generated content? Researchers at the University of Zurich have looked into this question.
How do AI-generated headlines resonate with people?
The one in the magazine PNAS Nexus The published study comes from the scientists Sacha Altay and Fabrizio Gilardi. You conduct research at the Institute of Political Science at the University of Zurich.
For their study on AI-generated headlines in news journalism, the two researchers conducted two online experiments with around 5,000 participants. These came from the USA and the UK.
The goal was to determine the impact of labeling headlines as AI-generated. The test subjects had to rate 16 headlines that were either true, false, generated by AI or by humans.
The details of the experiments
The test subjects were randomly given headings to evaluate for the study. These were divided into one of the following categories:
- no headline is marked as AI-generated
- AI-generated headlines are marked as such
- Human-generated headlines are marked as AI-generated
- false headlines are marked as false
In particular, the headlines marked as AI-generated did not perform well with the test subjects. These were rated as less accurate.
In addition, the test subjects were less willing to pass on articles with AI-generated headlines. The latter stood out regardless of whether the headlines were true or false. It was also irrelevant whether the headlines were actually generated by AI or were just labeled as such.
What is the reason for the aversion to AI-generated headlines?
The researchers also investigated the question of what causes the rejection of AI-generated headlines. To do this, they experimentally manipulated the definition of these AI-generated headlines.
The scientists were able to find out that the rejection of AI can be traced back to a specific circumstance. Because people who see content as AI-generated assume that this content was created entirely without human intervention.
According to the authors of the study, caution is therefore advised when labeling AI content. This labeling could unintentionally have negative consequences. Therefore, more transparency regarding the meaning behind these labels is necessary in the future.
Also interesting:
Source: https://www.basicthinking.de/blog/2024/11/11/ki-generierte-ueberschriften/
