< Back to situation

[REVISION HISTORY]

AI short stories rated higher than human

Updated 2 times since CLSTR started tracking revisions of this situation.

What changed

2026-08-07 05:44 UTC → 2026-08-09 18:33 UTC · added removed

Researchers at Villanova University have repeatedly found that short fiction generated by AI systems such as ChatGPT receives higher quality and engagement ratings than human‑written stories. In the original experiments, U.S. adults read six stories—three human and three AI—and, even when told the pieces were human‑authored, rated the AI works more favorably, revealing a bias toward perceived human authorship. Follow‑up studies showed that participants could not reliably identify the source, with correct identification rates between 39 % 39% and 52 %. 52%. A 2026 series of three experiments involving about 1,700 adult readers extended these findings. Half the stories were deliberately mislabeled, presenting AI‑generated texts as human‑written and vice‑versa. Regardless of the label, AI‑generated stories earned higher scores, and the label “written by a human” boosted ratings for both AI and human texts. Participants again struggled to discern origin, and researchers noted that AI narratives were perceived as clearer, more direct, and explicitly named topics, whereas human stories were subtler and more complex. The work, Further findings published in *Judgment and Decision Making*, underscores a gap between public assumptions about creative writing Making* by researchers Sydney Sears and the capabilities Deena Skolnick Weisberg highlight that while readers struggle to distinguish AI prose from established literary sources in blind tests, psychological bias remains significant: informing readers a story is human-authored increases its rating regardless of modern language models and highlights the need for greater actual origin. While AI literacy. can produce engaging and immersive narratives, researchers note that human perception remains heavily influenced by perceived authenticity, though they emphasize that AI's capabilities do not diminish the value of human creative self-expression.

Versions

  1. 2026-08-09 18:33 UTC AI short stories rated higher than human
  2. 2026-08-07 05:44 UTC AI short stories rated higher than human
  3. 2026-08-07 03:32 UTC AI short stories rated higher than human

Only revisions since CLSTR began indexing content versions appear here. Select a version to see what changed compared to the one before it.