started · updated
ChatGPT short stories outscore human writing in reader studies
A study published in the journal *Judgment and Decision Making* reveals that readers often rate ChatGPT-generated short stories higher than those written by humans in terms of quality and immersion. Researchers Sydney Sears and Deena Skolnick Weisberg conducted experiments involving thousands of participants to compare AI-generated prose with works from established literary sources.
In blind tests, participants struggled to distinguish between the two, with accuracy rates often near chance. However, the study highlights a significant psychological bias: when readers were informed a story was written by a human, they awarded it higher ratings regardless of its actual origin.
Findings suggest that while AI can produce clear and engaging narratives, human perception is heavily influenced by the perceived authenticity of the author. Deena Skolnick Weisberg noted that while AI can compete with human writing, it does not diminish the value of human writing as a form of creative self-expression.
Entities
ChatGPT · Deena Skolnick Weisberg · Deena Weisberg · Judgment and Decision Making · Sydney Sears · Villanova University
Claims
What the coverage asserts, and how many sources carry each claim.
- [● 4 SOURCES] Researchers concluded AI‑generated texts are clearer, more direct, and easier to process than human‑written stories. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] In the follow‑up experiments, fewer than half of participants could correctly distinguish AI from human stories. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] Two follow‑up experiments involved about 900 participants in total. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] Less than half of participants could correctly identify whether a story was AI- or human-generated. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] In the first experiment, half of the stories were mislabeled as human or AI. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] Approximately 1,700 participants took part in the first experiment. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] Participants who thought a story was written by a human rated it more positively. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de
- [● 4 SOURCES] The study was published in the journal “Judgment and Decision Making”. www.nordbayern.de · www.onetz.de · www.stimberg-zeitung.de · www.volksfreund.de