Readers tend to find news stories generated by artificial intelligence algorithms more credible, although they gain more pleasure from reading pieces written by human journalists, a study has found.
Researchers from the Ludwig-Maximilians-Universitaet in Munich, Germany, worked with 986 test subjects, which were tasked to read and evaluate a set of online news stories. Some of these stories, covering sports and business, were written by human journalists, while others were produced by computer algorithms.
Computer-generated stories are already being used by some of the world’s leading news agencies including Associated Press, as well as some German publishers.
In the study, the subjects were deliberately misled about who wrote each of the news stories. In multiple cases the researchers claimed that a story was written by a journalist when it was actually written by a robot.
The surprising finding was that stories attributed to human journalists were consistently rated as more readable and credible even though they were actually put together by robots.
A closer look at the data revealed that while stories actually written by humans were considered as more readable than computer-generated texts, the robot-produced pieces scored higher in credibility. The researchers are unsure of what created this impression.
"The automatically generated texts are full of facts and figures and the figures are listed to two decimal places. We believe that this impression of precision strongly contributes to the perception that they are more trustworthy," suggested Mario Haim, one of the authors of the paper.
There was a clear bias among the study participants when it came to evaluating readability. When told that a story was written by a human journalist, the tested subjects were inclined to rate it high on readability, even if it was actually written by the artificial intelligence system.
"To explain this finding, we assume that readers' expectations differ depending on whether they believe the text to have been written by a person or a machine, and that this preconception influences their perception of the text concerned," said Haim.
Computer programmes are used to generate simple news stories for sports and financial sections by putting together source data that is already structured in predictable ways.