Every day, we tell stories to make sense of life and to share our experiences. In the past, these stories were exchanged between humans, but with the rise of generative AI, a new conversational partner has emerged. Communication now also takes place between humans and machines.

“We know how important storytelling is in everyday life — for creating coherence and meaning, for influencing others, and for entertaining. At the same time, we know very little about this new form of interaction, which is beginning to spread on a large scale, where one can talk to a chatbot as if it were a human. But we have an idea that it will change the ways in which we use storytelling,” says Stefan Iversen, associate professor of Scandinavian Languages and Literature at Aarhus University.

In the new research project, he is working with assistant professor Ann-Katrine Schmidt Nielsen and postdoctoral researcher Pernille Meyer, both from Scandinavian Languages and Lliterature, to investigate the cultural and social effects when stories are created and shaped in interaction with machines.

The project focuses on narratives that arise when people use generative AI platforms such as ChatGPT, Gemini, and Replika for purposes beyond practical and task-oriented problem-solving.

“There is a strong focus on how AI is used as a tool in the workplace, but there is little research on non-professional use. This is one reason why we are curious to investigate it,” explains Ann-Katrine Schmidt Nielsen.

Examining the consequences of prompted conversations

In recent years, social media has influenced the ways in which we communicate with one another. According to the researchers, however, interactions with generative AI constitute a different form of communication than that which typically dominates social media.

“This is a new form of digital practice, which is more introverted. Its form is closer to that of a diary or a letter,” explains Stefan Iversen.

At the same time, it is increasingly possible to control the dialogue with language-model chatbots. Whereas on social media or elsewhere one may encounter resistance or disagreement, interactions with a chatbot can be tailored to specific needs and preferences.

“With the ability to personalize the chatbot, we can increasingly shape it in a particular direction. It is extremely interesting to examine what happens when the communication we engage in is largely controlled by ourselves,” says Ann-Katrine Schmidt Nielsen.

The research project will investigate the role of storytelling in three types of social interaction with chatbots. The researchers will examine the use of chatbots to create personal stories, such as when parents use prompting to generate bedtime stories featuring their child as the main character, rather than reading from existing children’s books. They will study the use of chatbots in the development of game worlds, for example in role-playing games and dynamic gaming environments. Finally, they will investigate practices in which users converse with a chatbot as if it were a person or a character. This may involve bereaved individuals communicating with a griefbot designed to imitate a deceased person, or with a fictional persona that can take on the characteristics of a bodiless life partner.

“Our particular focus is on investigating the stories that are created when people develop these deep, in some cases romantic-like, relationships with a chatbot,” explains Stefan Iversen.

The project Human-Machine Narration. How Generative Artificial Intelligence Transforms Recreational Storytelling has received a grant of 3.1 million DKK from the Independent Research Fund Denmark.

Contact

Stefan Iversen
Associate professor of Scandinavian Languages and Literature
School of Communication and Culture
Aarhus University
Mail: norsi@cc.au.dk
Phone: +4540878570

Ann-Katrine Schmidt Nielsen
Assistant professor, Scandinavian Languages and Literature
School of Communication and Culture
Aarhus University
Mail: noraksn@cc.au.dk