Readers Prefer AI-Generated Stories Over Human Writing, Study of 1,682 Adults Reveals

AI-Generated Stories Outperform Human Writing in Quality Ratings

A study published in Judgment and Decision Making by Cambridge University Press reveals that readers consistently rate AI-generated stories higher than human writing. Researchers Sydney Sears and Deena Weisberg from Villanova University recruited 1,682 adults aged 18 to 81 to evaluate short stories. Participants who read AI-generated short stories gave them an average quality score of 1.54 on a scale from -3 to 3, compared to just 0.97 for human-written stories1

1

. AI-generated tales also scored higher on absorption metrics, averaging 1.42 versus 1.0 for human writing1

1

.

The research design split participants into groups, with half reading human-authored stories from literary journals and half reading ChatGPT 4.0-generated content. Each group was further divided, with participants told either correctly or incorrectly whether their story was AI-generated or human-written. Stories received the highest ratings when participants believed AI-generated content had been written by humans, revealing what Deena Weisberg calls "a bias toward narratives written by real people"3

3

.

Readers Cannot Distinguish AI from Human Writing

Two additional experiments tested whether readers could identify AI-generated content. In the first test with 424 participants, only 39% correctly identified which stories were AI-generated—worse than random chance1

1

. A second test with 481 participants showed 52% accuracy, statistically no different from flipping a coin4

4

. The slight improvement between tests, conducted months apart, suggests society may be adapting to recognize AI output patterns.

Distinguishing AI from human writing proved equally difficult in both directions. Weisberg noted surprise at this finding: "Participants were equally likely to get the question wrong by saying that an AI-written story was written by a human and by saying that a human-written story was written by AI"4

4

. This suggests readers relied on incorrect assumptions about what characterizes AI-generated content versus human writing.

AI Literacy Emerges as Key Detection Factor

While literary expertise offered no advantage in identifying AI-generated stories, AI literacy proved significant. Self-reported AI familiarity increased correct identification odds by 14% in one experiment4

4

. When measured using the Artificial Intelligence Literacy Scale, a validated assessment tool, each one-point increase raised detection accuracy by 33%4

4

. Weisberg explained that "familiarity with AI systems appeared to help people recognize the patterns typical of AI-generated writing, such as em dashes and sentence structures such as 'it's not just X, it's Y'"4

4

.

Why AI Writing Resonates with Readers

AI writing tends to be clearer, more direct, and easier to process than human-authored fiction. An AI-generated excerpt from the study demonstrates this simplicity: "So now, as I sit by the pond, the autumn leaves falling gently around me, I think of my mother, of the stories she shared, of the wisdom she imparted"2

2

. In contrast, human writing employs more complex literary devices and subtlety, requiring greater cognitive effort from readers.

Weisberg describes AI output as more "vanilla" content, attributing the quality preference to this characteristic1

1

. The preference reflects how AI mimics human tropes, syntax, and styles that audiences already favor. Since these systems train on decades of human-written content, they naturally incorporate familiar patterns. Rodney Jones from the University of Reading suggests the study's use of literary journal stories—rather than popular genres—may have amplified this effect: "People like expected things. People like easy-to-process things. Real human writing, real kind of human literary writing, is, frankly, not very popular, because it's hard work"1

1

.

Implications for Creative Fields and Authenticity

The findings align with a 2023 MIT study led by Renee Richardson Gosline, which found AI-generated content rated higher than work by human experts2

2

. Weisberg argues these results show "public assumptions about AI's capabilities are increasingly out of date"3

3

. Despite this, she maintains that human creativity retains unique value: "Humans write as a form of creative self-expression or to challenge ourselves, or to make sense of our experiences. The fact that AI can generate human-like stories doesn't change any of that"5

5

.

Ethical concerns surround AI-generated art and literature. Luke Kennard from the University of Birmingham emphasizes that AI represents "an existential threat" rather than a neutral tool, noting it relies on "a massive stolen database of actual writing" with "catastrophic repercussions for the surrounding communities"5

5

. These controversies mirror broader debates in creative fields, including a 2024 Commonwealth prize controversy where a short story published in Granta faced inconclusive investigation over AI generation allegations, and artist criticism of Christie's AI-generated art auctions as "mass theft"1

1

.

Weisberg plans future experiments testing different writing types to determine whether AI maintains its advantage across genres. She speculates AI will achieve artistic value measured on different scales than human creativity1

1

. As AI capabilities advance rapidly, improving AI literacy may help audiences navigate this evolving landscape and make informed judgments about content authenticity and story quality.