According to a new study, readers rate stories written by artificial intelligence higher in quality and content than human-created texts. As part of an experiment published in the journal Judgment and Decision Making, more than 1,600 adults read texts generated by both real authors and ChatGPT. The results showed that people struggle to distinguish technology from human writing.
The study revealed an interesting paradox: when readers did not know the true identity of the author, they gave the highest ratings to AI-generated texts because they were easier to read and understand. However, as soon as participants were told that a story was written by a human, they immediately rated it higher.
In additional experiments, scientists attempted to determine whether people could guess the author’s identity. It was found that the majority of readers lack the ability to accurately differentiate between works created by humans and AI, with the detection rate being equivalent to chance. Although the quality of technology-generated texts equals or exceeds that of humans, the study’s authors explain that this does not imply a decline in human writing skills.
According to experts, writing is a form of self-expression, and the development of artificial intelligence does not threaten this process; however, ongoing progress already requires establishing a place in the literary space for AI-generated or co-authored works. On the other hand, some specialists strongly criticize the use of artificial intelligence in writing.
