Does AI Authorship Undermine Reader Engagement?

Imagine reading a compelling short story—now imagine being told it was written by AI. Would it feel the same? This question lies at the heart of two recent experiments investigating AI authorship and its impact on narrative transportation (Messingschlager & Appel, 2024), the psychological immersion readers experience while engaging with a story.

Study overview

As AI-generated content becomes increasingly prevalent, understanding how readers emotionally respond to AI-written stories has become more crucial than ever. In the first experiment of this investigation, Messingschlager and Appel (2024) examined how the belief that a story was authored by either an AI or a human influenced readers’ experiences. In Experiment 1, participants were randomly assigned to read a brief author biography indicating either a human or AI author, followed by a contemporary fiction story—all of which were actually written by humans. After reading, participants rated: 1) Narrative transportation; 2) Eeriness (a sense of discomfort or uncanniness); 3) Willingness to share the story with others; 4) Openness to experience and affinity for technology. Experiment 2 extended this design by incorporating both contemporary fiction and science fiction stories to examine whether genre influences the impact of perceived AI authorship.

Findings

In Experiment 1, the results revealed that narrative transportation significantly declined when readers believed the story was authored by AI. Interestingly, and contrary to expectations, AI authorship made the story feel less eerie, rather than more. Additionally, participants’ willingness to share the story was unaffected by the perceived authorship. Individual differences (i.e., openness to experience and technological affinity) did not moderate the effects of AI authorship on transportation, eeriness, or sharing intentions. While greater openness was associated with higher transportation and willingness to share overall, it did not buffer against the reduced engagement caused by AI attribution. 

In Experiment 2, the main finding was replicated: for contemporary fiction, perceived AI authorship again reduced narrative transportation. However, when it came to science fiction, no significant difference emerged between the AI and human author conditions. This suggests that genre plays a crucial role. Readers may view AI as a more legitimate or credible author within the science fiction genre—possibly because of the genre’s inherent association with technology and futurism. In this context, AI authorship may feel more fitting or even advantageous.

Disclaimer

This blog post is based on scientific research and is intended solely for public awareness.

For detailed information, please refer to the original article.

Messingschlager, T. V., & Appel, M. (2024). Creative artificial intelligence and narrative transportation. Psychology of Aesthetics, Creativity, and the Arts, 18(5), 848-857. https://doi.org/10.1037/aca0000495 

What Happens in Our Brain When We Read Literary Short Stories?

Have you ever wondered what happens in your brain when you lose yourself in a gripping short story? Research shows that reading fiction isn’t just a passive activity—it’s an immersive experience that engages multiple brain regions. When we read, we simulate the story in three distinct ways: motor simulation (imagining physical actions), perceptual simulation (re-enacting sensory experiences), and mental simulation (understanding characters’ inner worlds). But what are the neural substrates of these simulations? Do they follow distinct pathways, or do they converge in a specific brain region? Mak and colleagues' (2023) study sheds light on these questions.

Study overview

In this study, researchers recruited 40 participants aged 18–43 with normal vision and no history of dyslexia, brain surgery, or epilepsy. Participants read two short, descriptive stories inside an MRI scanner and equipped with an eye-tracking device to monitor their eye movements during reading. While still undergoing brain scanning, their subjective experiences of the stories were assessed using questionnaires measuring story world absorption and story appreciation. After the scan, their trait-level personal characteristics were evaluated through empathy and transportability questionnaires.

Findings

The study showed three distinct but overlapping neural pathways activated during narrative processing: motor simulation (premotor cortex and supplementary motor area), perceptual simulation (visual cortex and sensory regions), and mental simulation (temporoparietal junction and medial prefrontal cortex). These pathways converge in the supramarginal gyrus*, an important region for referential indexing (integrating references within a text), situation model building, and integration of sensory, emotional, and cognitive aspects of a story.

Eye movement analysis showed that motor descriptions were read faster, while perceptual and mental descriptions slowed reading due to more complex cognitive demands. Furthermore, trait-level factors like empathy and transportability had a stronger influence on brain activation than state-based factors like story absorption, emphasizing the role of personality in narrative simulation.

* The supramarginal gyrus functions as a hub where multiple brain networks intersect, including the default mode network (involved in self-perception and mind-wandering) and the mentalizing network (responsible for social cognition).

Disclaimer

This blog post is based on scientific research about the brain's response to reading fiction and is intended solely for public awareness.

For detailed information, please refer to the original article.

Mak, M., Faber, M., & Willems, R. M. (2023). Different kinds of simulation during literary reading: Insights from a combined fMRI and eye-tracking study. Cortex, 162, 115-135. https://doi.org/https://doi.org/10.1016/j.cortex.2023.01.014

Picture Credit.

Miss Auras, The Red Book by John Lavery