Study finds readers prefer AI-written stories to human ones

Readers gave the highest ratings to short stories generated by artificial intelligence — particularly when they believed the work had been written by a human — according to research published by Cambridge University Press in the journal Judgement and Decision Making.bbc+1

The study, led by researchers at Villanova University, tested whether participants aged 18 to 81 could distinguish between human-written and AI-generated fiction. They could not. Identification accuracy was just 39.93 percent in one experiment — worse than chance — and 51.97 percent in another, no different from a coin flip.eurekalert+1

How the Study Worked

Researchers used three short stories written by published authors alongside three generated by ChatGPT. In the first experiment, 1,682 participants read a story and were either correctly or incorrectly told whether it was written by AI or a human, then rated how engaging it was. In subsequent experiments, 424 and 481 participants received one human-written and one AI-written story without labels and attempted to identify which was which.bbc

"Participants gave the highest ratings to stories that they had been told were written by humans but which were actually written by AI," said senior author Dr. Deena Weisberg of Villanova's department of psychological and brain sciences. "I'd argue that this reveals a bias towards narratives written by real people."bbc

AI-generated stories also received higher quality ratings overall. "AI writing tends to be clearer, more direct and easier to process," Weisberg said. "By contrast, human-written stories are often more subtle and complex. Perhaps people generally prefer predictability, because difficult or subtle material requires more brain power."bbc

Implications for Education

The findings arrive as universities grapple with AI's role in academic writing. A growing number of institutions — including Yale, Vanderbilt, Johns Hopkins and Indiana Universities — have enacted policies banning or discouraging faculty reliance on AI-detection programs as sole evidence of cheating, according to Inside Higher Ed. At least a dozen universities, including Northwestern and Georgetown, have disabled Turnitin's AI detection software after it showed higher false-positive rates than originally claimed.insidehighered

"Over the past 20-plus years, there have been changes that have slowly degraded the validity of our 20th-century assessments," said Tricia Bertram Gallant, director of the academic integrity office at UC San Diego. "We're still relying on the unsupervised written word as evidence of learning. The real trick is acknowledging that that doesn't work anymore."insidehighered

A Shifting Landscape

Weisberg noted that public assumptions about AI's creative capabilities are "increasingly out of date". The study adds to a growing body of research challenging assumptions that creative writing requires uniquely human qualities such as emotional understanding and lived experience. People who reported greater AI literacy were better able to identify AI-written text, but those who claimed expertise in literature were not.bbc